320 defines a second door area 340. The door areas are in front of the automatic door. While the fields of view are shown as rectangular, it may be understood that the door areas can be any shape and may not match in shape or area. The first door area 330 and the second door area 340 can overlap in an overlap region 350. A target in the overlap region is detected by each radar sensor. Each radar sensor is configured to capture radar data based on a coordinate system defined by the radar sensor. For example, a target located in the overlap region may appear in a left or right portion of the field of view depending on the radar sensor. The radar data from each radar sensor can be combined when the relative position/orientation of the radar sensors is known. For example, as shown in FIG. 3, the second radar sensor 320 is spatially offset from the first radar sensor in the X-direction of the coordinate system (X, Z) of the first radar sensor 310. If this offset is known, the radar data from the second radar sensor 320 can be registered to the coordinate system of the first radar sensor 310. The radar data can then be combined so that the radar data from the two sensors forms a combined door area that is the union of the first door area 330 and the second door area 340.

The implementation shown in FIG. 3 includes two radar sensors. In general, the door sensor system may include any number of radar sensors. A plurality of radar sensors can communicate with each other and/or with the DCU 210 in a variety of ways. FIGS. 4A-4C are block diagrams showing various communication configurations for the plurality of radar sensors. FIG. 4A illustrates a first communication configuration in which each radar sensor 400 (i.e., S) communicates directly with the DCU 210. The direct communication configuration may eliminate the need for a communication protocol to communicate on a common channel (i.e., bus) but may add complexity to the DCU. FIG. 4B illustrates a second communication configuration in which each radar sensor 400 and the DCU 210 communicates to share information over a bus 410. In some implementations, the DCU 210 may require a hub 420 to communicate with the bus 410. FIG. 4C illustrates a third communication configuration in which the radar sensors form a sensor network 430 to share information between each radar sensor 400. One of the radar sensors can be designated/configured as the master sensor 435. The master sensor 435 can be configured to manage the communication on the sensor network. Additionally, the master sensor 435 may function as an interface with the DCU 210. The sensor network 430 can be wired or wireless.

FIG. 5 is a block diagram of a radar sensor according to a possible implementation of the present disclosure. The radar sensor 400 includes a mmW radar module (i.e., mmW radar 600). The mmW radar 600 is configured to detect a position in one, two, or three dimensions (e.g., spherical position, cartesian position) of a target and the movement (e.g., path, velocity, acceleration) of the target.

The radar sensor 400 further includes an inertial measurement unit (i.e., IMU 510). The IMU 510 includes a combination of sensors to measure the orientation of the IMU (e.g., with respect to the earth). The combination of sensors can include accelerometers, a gravitometer, gyroscopes, and/or a magnetometer. The IMU is mechanically coupled (to the mmW radar 600 so that IMU data returned by the IMU 510 corresponds to the orientation of the mmW radar 600.

The radar sensor 400 further includes a sensor control unit (SCU 520). The SCU 520 can be configured to communicate radar data and IMU data with the DCU and/or with other sensors, as described above. In a possible implementation, the SCU 520 can be configured to determine the orientation of the mmW radar from the IMU data and register the mmW radar data based on the determined orientation. The SCU 520 may include memory 530 on which to store (e.g., buffer) radar/IMU data and/or software for radar/IMU data processing.

FIG. 6 is a block diagram of a mmW radar according to a possible implementation of the present disclosure. The mmW radar 600 includes a processor 601 that is coupled to a transmitter 610 and a receiver 620. The mmW radar 600 may include multiple receivers in order to determine a direction of a target based on a phase difference between signals received at each antenna 621 of the receivers. The multiple receivers may be arrayed in two-dimensions in order to determine a phase shift in a first dimension and a phase shift in a second dimension. The two dimensions may correspond to two angles that define a target's position in two dimensions, which can be combined with range to locate the target in three dimensions. A resolution of an angle (i.e., angular resolution) in a particular dimension may correspond to the number of receivers in the particular dimension.

The configuration shown in FIG. 6 includes a single transmitter and a plurality of receivers (i.e., receiver array) spaced (e.g., equally spaced) along a direction of the receiver array. This configuration is known as a single-input-multiple-output (SIMO) radar. It is also possible to have a multiple-input-multiple-output (MIMO) radar that includes multiple transmitters. MIMO antennas may accomplish higher angular resolutions with fewer receiver modules. The principles of the present disclosure can be applied to either of these configurations. For the purposes of explanation, the SIMO configuration shown in FIG. 6 will be discussed in detail.

The transmitter 610 includes a synthesizer to produce a transmit signal (i.e., Tx) that changes frequency over time according to a ramp profile. For example, the transmitter 610 can include a ramp generator 611 to repeatedly generate a voltage that steadily increases (or decreases) with time for a period. The changing (i.e., ramp) voltage may then be coupled to an input of a voltage-controlled oscillator (i.e., VCO 612) or other oscillator that has a mmW frequency that is proportional to the input voltage. The resulting frequency modulated (i.e., chirped) transmit signal (i.e., Tx) is then coupled to a field of view via a transmit antenna 613. The mmW can be referred to as a frequency modulated continuous wave (FMCW) radar.

Each receiver 620 of the mmW radar 600 includes an antenna 621 configured to receive signals reflected from an object in the field of view. A range to an object can cause a delay between the transmitted signal (i.e., Tx) and the received signal (i.e., Rx), which for a chirped signal, corresponds to a frequency difference. As a result, the processor 601 of the mmW radar 600 can estimate distance to a target (i.e., range) based on the frequency difference between the transmitted signal and the received signal. Accordingly, each receiver of the mmW radar 600 further includes a mixer 622 configured to determine the frequency difference between the received signal (Rx) and the transmit signal (Tx). Each receiver may further include a filter 623 (e.g., low-pass filter, band-pass filter). The filter may help to isolate an intermediate frequency (i.e., IF) signal from harmonics resulting from the mixing process. The IF signal can include multiple tones, where the frequency of each tone is proportional to the distance (i.e., range) of a corresponding target (i.e., object). Each receiver may further include an analog to digital converter (i.e., ADC 624) that is configured to digitize the IF signal. The digitized signal may then be coupled to the processor 601 where it can be processed (e.g., transformed) in order to generate the mmW radar data which includes the range information to each target. For example, a Fourier transform (e.g., fast Fourier transform (FFT)) may be performed on the ADC data to obtain a frequency domain representation (i.e., spectrum) of the IF signal. The peaks of the spectrum can correspond to the range of the objects. Accordingly, the processor may be configured to identify range of each target based on peak detection of a spectrum. The mmW radar data may include a range to an object in the field of view.

The mmW radar data may further include an angle 635 of the received signal from the object with respect to the mmW radar. A coordinate system 630 of the mmW radar may be based on a direction of the antenna array of the receiver and the angle 635 may be relative to an axis of the coordinate system. Because of the angle 635, the received signal at each of the receivers will be shifted in phase with respect to each other. In other words, the range from the object to each of the receiver antennas is slightly different. To determine the angle, the processor 601 may be configured to determine the phase of corresponding spectral peaks at each receiver to determine the phase shifts between antennas. In other words, the processor may be configured to determine how a phase of a spectral peak changes over position to determine a spatial phase shift. The processor 601 may then be configured to relate the spatial phase shift to an angle to the target.

The processor 601 may be further configured to combine the range and the angle of mmW radar data to determine the position (i.e., location) of an object in the field of view (i.e., door area). When multiple objects/entities are in the field of view the processor 601 may be configured to determine positions of each. Multiple object positions may be resolved based on the angular resolution of the mmW radar. The angular resolution can correspond to the number of transmitters/receivers in the mmW radar.

The mmW radar data may further include a velocity measured for a moving object in the field of view. In order to measure velocity (i.e., speed), the processor 601 may be configured to estimate how the phase shift of a spectral peak changes over time. In other words, the processor may be configured to determine how a phase of a spectral peak changes over time to determine a temporal phase shift. The temporal phase shift may correspond to the velocity. Objects at the same range (i.e., having the same spectral peak) can be resolved using a frame of chirp signals and then performing a second transform (i.e., doppler-FFT) to resolve the objects by peaks corresponding to their different temporal phase shifts (i.e., velocity).

FIG. 7A illustrates the coordinate system 630 of a mmW radar. An object (i.e., target) may be located in space based on the coordinate system 630, which can be physically centered at the mmW radar 600 and aligned with the physical axes of the mmW radar 600. Accordingly, the mmW radar data may include the position of a target 660 expressed in three dimensions. For example, a position of the target relative to the mmW radar may be expressed as a radial dimension (i.e., R), an azimuthal angle (i.e., ϕ), and an elevation angle (i.e., θ). Alternatively, the position of the target 660 may be represented (e.g., converted from spherical coordinates) as cartesian coordinates (X, Y, Z). In some implementations, position of the target 660 may be used to determine (i.e., mapped to) a position in a door area and/or to a zone in the door area. In some implementations, the position of the target 660 may be used to determine a height of the object off the floor. For example, a first range to the floor may be computed and a first range to the object may be computed and a height of the object may be computed based on a difference between the first and second range.

Each mmW radar may generate mmW radar data based on a coordinate system corresponding to the mmW radar. In other words, each mmW radar may generate mmW radar data relative to its own coordinate system. In an automatic door system, it may be desirable to use multiple mmW radars to expand a sensed area. Each coordinate system of the multiple mmW radars may be different. To compared and/or combined mmW radar data from different mmW radars, it may be necessary to convert the mmW radar data from the mmW radars to a common (i.e., reference) coordinate system. For example, mmW radar data from each mmW radar may be converted to a reference coordinate system (e.g., corresponding to the door). In another possible implementation mmW radar data from a first mmW radar may be converted to the coordinate system of a second mmW radar. Comparing and/or combining mmW radar data from multiple mmW radars may include a coordinate transformation (i.e., conversion) of the mmW radar data to a common (i.e., reference) coordinate system. The converted mmW radar data may be referred to as registered mmW radar data.

FIG. 7B illustrates how coordinate systems can be compared for a coordinate conversion. A transformation between a first coordinate system 650 and a second coordinate system may include as a rotation about the first coordinate system z-direction (i.e., yaw), a rotation about the first coordinate system's y-direction (i.e., pitch), and a rotation about the first coordinate system's x-direction (i.e., roll). In other words, if a yaw, pitch, and roll of a coordinate system relative to a reference coordinate system is known, then the data related to range, angles, and positions of the coordinate system may be transformed (i.e., converted) to the reference coordinate system. Accordingly, a conversion may include determining a yaw, a pitch, and a roll between first coordinate system and a second coordinate system. Because IMU data corresponds to a coordinate system, the conversion between radar sensors may include determining a yaw, a pitch, and a roll based on a comparison of IMU data from the radar sensors. For example, a first IMU data from a first radar sensor and a second IMU data from a second radar sensor may be compared to determine a conversion necessary to register the mmW radar data from the first radar sensor. In this case, the position of one of the radar sensors may be the reference coordinate system. In some cases, additional information is necessary for a conversion. For example, a reference coordinate system may be aligned with the door and a conversion may require the details of the door as they relate to the reference coordinate system. Further, a conversion may include determining an offset between coordinate systems. For example, an offset may correspond to a spatial separation of radar sensors in the door frame.

Returning to FIG. 5, each radar sensor may include an IMU that is mechanically coupled to the mmW radar 600. The IMU is configured to generate IMU data based on the orientation of the mmW radar relative to a reference coordinate system. The reference coordinate system may be an earth coordinate system, a door coordinate system or a coordinate system of another mmW radar. The SCU 520 may include a processor that can be configured to receive the mmW radar data and the IMU data and to convert the mmW radar data to the reference coordinate system based on the IMU data. The conversion may also include knowledge of a physical layout. In other words, registering the mmW radar may require knowledge of both orientation and position (i.e., location) relative to a reference coordinate system.

FIG. 8 illustrates a possible physical layout of a first mmW radar 801 and a second mmW radar 802. Each mmW radar has its own local coordinate system. Each local coordinate system is rotated relative to a reference coordinate system 800. In particular, a first coordinate system 811 for the first mmW radar 801 is rotated −θ with respect to the reference coordinate system 800 and a second coordinate system 812 is rotate +θ with respect to the reference coordinate system 800. In addition, each mmW is offset relative to the reference coordinate system. In particular, the first coordinate system 811 is offset +d with respect to the reference coordinate system 800 and the second coordinate system 812 is offset −d with respect to the reference coordinate system (i.e., the second coordinate system 812). The offset may be determined by the frame of the door without the need for active (e.g., IMU) measurement. When relative orientation and location of the mmW radars is known, then a position of an object 820 may be determined with respect to any of the coordinate systems based on data from either the first mmW radar 801 or the second mmW radar. The reference coordinate system may be a coordinate system of the door or a coordinate system of one of the mmW radars.

When mmW radar data is converted to the reference coordinate system (i.e., common coordinate system) it is said to be registered. For example, mmW radar data from a first mmW may be registered to a door coordinate system and mmW data from a second mmW radar may be registered to the door coordinate system. The registered mmW radar data can be combined to formed combined radar data. The combined radar data can be analyzed to locate a target in a combined door area.

Registering first mmW radar data (from a first mmW radar) and second mmW radar data (from a second mmW radar) to a reference coordinate system can include determining a first coordinate conversion between a first coordinate system (of the first mmW radar) and the reference coordinate system and a second coordinate conversion between a second coordinate system (of the second mmW radar) and the reference coordinate system. Next, the first mmW radar data may be modified according to the first coordinate conversion to obtain registered first mmW radar data, and the second mmW radar data may be modified according to the second coordinate conversion to obtain registered second mmW radar data. The registered first mmW radar data and the registered second mmW radar data can then be combined to obtain combined mmW radar data.

The reference coordinate system may be selected variously. In a possible implementation the reference coordinate system is a coordinate system aligned with the physical shape of the automatic door. For example, a horizontal axis (e.g., x-axis) of a reference coordinate system can be in a direction along which the door panels open and close (i.e., door opening 103), and a vertical axis (e.g., y-axis) of the reference coordinate system can be along a direction from the threshold 115 to the header 110 of the sliding door. In another possible implementation, the reference coordinate system is a coordinate system of one of the mmW radars (e.g., the master sensor). In this implementation, mmW radar data from other radar sensors are converted to the coordinate system of the master sensor. This conversion may be aided by communication of IMU data amongst radar sensors.

FIG. 9 is a flowchart of a method for generating combined mmW radar data 940 according to a first possible implementation of the present disclosure. The implementation shown can be for a door sensor system of an automatic door system that includes (i) a first radar sensor (i.e., S_1) including a first IMU (i.e., IMU_1), a first mmW radar (i.e., mmW radar_1) and a first sensor control unit (i.e., SCU_1) and (ii) a second radar sensor (i.e., S_2) including a second IMU (i.e., IMU_2), a second mmW radar (i.e., mmW radar_2) and a second sensor control unit (i.e., SCU_2). The dual radar sensor implementation is presented for simplicity, but the method could be expanded to include any number of radar sensors.

The method 900 includes collecting 911, at the first radar sensor (S_1), IMU data using IMU_1 and collecting 912 mmW radar data using mmW_radar_1. The method 900 further includes collecting 921, at the second radar sensor (S_2), IMU data using IMU_2 and collecting 922 mmW radar data using mmW_radar_2. The data from each IMU and each mmW radar is then communicated to a door control unit (i.e., DCU) of the automatic door system. For example, SCU_1 and SCU_2 may each be configured to communicate the data over one of the topologies described previously (see FIGS. 4A-4C).

The method 900 further includes determining 931 a conversion for the radar data (e.g., range, angle, position, velocity) based on the IMU_1 data and the IMU_2 data. For example, he DCU 930 may be configured to receive the IMU data, determine the relative orientation and positions of S_1 and S_2 based on the IMU data, and determine the data conversion based on the relative orientations and positions. Based on the determined conversion, the method 900 then includes registering 932 the mmW radar data to a common coordinate system. For example, the DCU may be configured to register the mmW radar data to a reference coordinate system (e.g., automatic door coordinate system) and combine the mmW radar data into combined mmW radar data 940. This method 900 for generating the combined mmW radar data may have the advantage of performing more complex algorithms for data registration at the DCU which can centralize the processing requirements, which may provide cost, complexity and power savings.

FIG. 10 is a flowchart of a method for generating combined mmW radar data 1040 according to a second possible implementation of the present disclosure. The method 1000 includes collecting 1011, at the first radar sensor (S_1), IMU data using IMU_1 and collecting 1012 mmW radar data using mmW_radar_1. The method 1000 further includes collecting 1021, at the second radar sensor (S_2), IMU data using IMU_2 and collecting 1022 mmW radar data using mmW_radar_2. In this implementation IMU data may be communicated between the radar sensors, such as by one of the communication topologies described previously (e.g., see FIGS. 4A-4C). This communication can enable the registration to take place using the processing (i.e., the SCU) of each radar sensor.

The method 1000 further includes generating registered mmW radar data at the first radar sensor (S_1) and the second radar sensor (S1). The method 1000 further includes determining 1013 the conversion and registering 1014 the first mmW radar data (i.e., mmW_radar_1). The first SCU (i.e., SCU_1) can be configured to determine the conversion based on the first IMU data (IMU_1) and second IMU data (IMU_2) received form the second radar sensor (S_2). For example, the relative orientation of the coordinate systems for each radar sensor may be calculated to determine the orientation. The method further includes determining 1023 the conversion and registering 1024 the second mmW radar data (i.e., mmW_radar_2). For example, the second SCU (i.e., SCU_2) can be configured to determine the conversion based on the second IMU data (IMU_2) and first IMU data (IMU_1) received from the first radar sensor (S_1).

The method 1000 further includes communicating the registered mmW radar data to the DCU of the automatic door system. The communication may be accomplished according to one of the topologies described previously, such as shown in FIGS. 4A-4C. The method 1000 further includes combining 1032 the registered radar data at the DCU to generate combined mmW radar data 1040. This method 1000 for generating the combined mmW radar data may have the advantage of versatility because only registered data is communicated. Accordingly, radar sensors may be easily added or removed from the automatic door system.

The conversions described may include selecting a reference coordinate system upon which to compare the orientation of a radar sensor (i.e., IMU data). The reference coordinate system may be a fixed coordinate system, such as described by the dimensions of an automatic door. In this case, no IMU data may be necessarily collected for the door because its orientation and position are fixed and known. Alternatively, the reference coordinate system may be a coordinate system of a particular radar sensor (e.g., master sensor). In this case, the IMU data from the particular radar sensor may be collected and used as a reference to which the IMU data from other radar sensors may be compared to determine the necessary conversion for registration. Variations to these approaches may exists and are within the scope of the present disclosure.

FIG. 11 illustrates a combined door area generated by two radar sensors according to a possible implementation of the present disclosure. The data from the first radar sensor and the mmW radar data from the second radar sensor are registered so that the positions in the two door areas both reference a common origin. A first door area 1101 has N×M possible positions and includes a starting position (0,0) that is the origin. A second door area 1102 also has N×M possible positions but is spatially offset by from the origin in an x-direction and a y-direction. Accordingly, the starting position for the second door area is (+X₁, −Y₁). The door areas include an overlap region (i.e., shown as hashed) which is included in the fields of view (door areas) of both radar sensors. For example, a target may be found at position (N, M) in registered radar data from both the first radar a sensor and the second radar sensor.

The first door area 1101 and the second door area 1102 can be included in a combined door area including combined mmW radar data. The combined mmW radar data may be the union of radar data from the first radar sensor and radar data from the second sensor. The combined door area may include a region including radar data from the first radar sensor (and not the second radar sensor) and a region including radar data from the second sensor (and not the first radar sensor). In the overlap region, the registered radar data from each sensor should ideally match. For example, a target located at a position in the overlap region should be measured at the same position by each radar sensor after the data is registered. Accordingly, combining radar data in the overlap region may include selecting one of the radar sensor registered data or processing (e.g., averaging) the registered data from the two sensors for each point in the overlap region. Zones can be defined in the combined door area so that when a target is detected at a position in the combined door area that corresponds to a zone, a corresponding action (or inaction) may be triggered. Further key positions may be defined in the combined door area, such as a center of the door.

The combined mmW radar data can be a three-dimensional array (i.e., point cloud) of received (i.e., returned) signals from the radar sensors. Each point in the point cloud can be associated with a range to a target from the door and an angle to the target with respect to the door. A strong received signal at a point may indicate a target located at the point. A target may include a group of high received signal points that are clustered in a shape corresponding to the target. In other words, multiple points in the point cloud may correspond to a target. A range to the target may be determined from the ranges of the points in the point cloud corresponding to the target and an angle of the target may be determined from the angles of the points in the point cloud corresponding to the target. The ranges and the angles of the target can change as the target moves. The combined mmW radar data may be analyzed to track one or more targets in a combined door area corresponding to the combined mmW radar data.

A moving target can be tracked through the combined area. For example, the target at a first time is determined to be at a first position 1110 that is included in the first door area 1101, the target at the second time is at a second position 1120 that is included in both the first door area and the second door area, and the target at a third time is at a third position 1130 that is included in the second door area. Tracking may include determining which door area (i.e., radar sensor) includes the target and handing off the tracking between door areas (i.e., radar sensors) as the target moves.

The IMU data for each radar sensor 400 may be expected to remain the same after a service process, such as installation, commissioning, calibration, or re-calibration, is performed. Accordingly, each radar sensor 400 may be configured to monitor its position over time after the service process. which may be useful in generating a notification for a new service process, such as re-calibration. FIG. 12 is a flowchart of a method 1200 for detecting a change in position of a radar sensor. The SCU 520 and the memory 530 of each radar sensor in a door sensor system may be configured by software to perform one or more of the steps of the method 1200.

In the method 1200, baseline IMU data is captured 1210. The capture of the baseline IMU may be triggered manually. For example, a service technician may trigger an SCU 520 to capture and store baseline IMU data in a memory 530 as a baseline orientation 1260. The capture of the baseline IMU data may also be triggered automatically. For example, the SCU 520 may be configured to capture and store a baseline orientation 1260 (e.g., in a memory 530 of the SCU) periodically (e.g., quarterly, yearly) to account for drifting measurement results due to external factors, such as age and temperature. After storing the baseline orientation (i.e., baseline IMU data), the method 1200 may include capturing 1220 IMU data. For example, the SCU 520 may be configured to receive the IMU data captured by the IMU regularly as part of the target sensing for the automatic door operation. The method 1200 may further include comparing 1230 the baseline IMU data (i.e., baseline orientation) and the IMU data to determine if a change 1240 in orientation of the mmW radar has occurred. For example, the SCU 520 may be configured to receive IMU data from the IMU and recall baseline IMU data (i.e., baseline orientation) from a memory 530 of the SCU and perform a comparison. If the comparison satisfies a criterion (e.g., exceeds a threshold) then the SCU may be configured to transmit 1250 a signal based on the change. The method may repeat the comparison (e.g., regularly, as needed, etc.) in order to monitor the position/orientation of the mmW radar.

The signal may be used to trigger events such as a service alert, a recalibration (e.g., automatic recalibration), or the like. In a possible implementation, a change in the position and/or orientation of the mmW radar 600 can trigger the sensor control unit (i.e., SCU 520) to transmit a signal to the DCU 210 to indicate that the IMU data for a radar sensor in the door sensor system 220 has changed. This signal may trigger the DCU 210 to recalibrate the door sensor system and/or to transmit an alert via the door interface 260. For example, a possible alert can be a notification that service of the automatic door system 200 is required. In another example, the signal may trigger the DCU to determine and update values used for the conversion to register and combine mmW radar data.

The combined mmW radar data can include more information than required simply to detect motion and presence. Control of the door can be controlled according to more conditions (i.e., enhanced) based on the additional information. The enhanced door control can be tailored to match a target and its movement and can be less prone to errors (e.g., false opening/closing). Further, this control can account for multiple targets and can adapt to changing conditions. This may have the technical effect of increasing the safety and reliability of an automatic door.

Returning to FIG. 2, the combined mmW radar data may be processed by the DCU 210 to generate signals that control the opening and closing of the door. For example, based on an analysis of the combined mmW radar data, the DCU 210 can be configured to control the operation of the door. The operation of the door may include an opening or a closing. Further, the operation of the door may include an amount, a position, or a timing of the opening/closing. The DCU 210 may control the operation by transmitting signals (e.g., pulse width modulated (PWM) signals, torque signals) to the door operator 230 which can translate these signals and operate a motor of the door assembly accordingly.

FIG. 13. is a flow chart illustrating functions (i.e., modules) that may configure the DCU for door operation. In other words, the modules may correspond to programs, subroutines, etc. running on a processor (or processors) of the DCU to configure the DCU to perform the function of the module. The DCU may include a target tracking module 1400. The target tracking module 1400 can be configured to receive the combined mmW radar data 1300 and to identify the one or more targets from the combined mmW radar data 1300. The target tracking module 1400 may be further configured to determine a position of each of the one or more identified targets, and the target tracking module 1400 may be further configured to determine movement of the one or more identified targets, which are in motion.

The results of the target tracking module can be transmitted to a door control module 1510. The door control module 1510 can be configured to process the information about the one or more targets to determine how, and when, the door should be controlled. Additionally, the results of the target tracking module 1400 can be transmitted to a door filter module 1530. The door filter module 1530 can be configured to process the information about the one or more targets and to identify certain conditions that could trigger certain door operation (or prevent certain door operations). For example, the door operations triggered by the door filter module 1530 may be associated with an operating requirement for safety and/or may be associated with an operating requirement for improving the reliability of the door. For example, the door filter module may be configured to prevent the automatic door from closing on a person in the doorway or from opening in response to movements that should not open the door.

Information from the door control module 1510 and the door filter module 1530 can be received by a door operation module 1550. The door operation module 1550 can be configured to determine when and how to open/close the automatic door based on the information from the door control module 1510 and the door filter module 1530. Some determinations made by the door operation module can include an amount of a door opening/closing, a position of the door opening/closing, a speed of the door opening/closing, and/or a time of the door opening/closing. The door operation module 1550 may be further configured to translate the details regarding the door opening/closing into signals that can be transmitted from the DCU to the door assembly (e.g., the door operator) for door control. Information from the door assembly may be fed back to the door control and/or door filter to help operate the door. For example, the determination of how and when to open/close the automatic door may be based on the state (e.g., opening, opened, closing, closed, etc.) of the door assembly.

FIG. 14 is a flowchart illustrating details of the target tracking module 1400. In other words, FIG. 14 is a flowchart illustrating a method for target tracking according to a possible implementation of the present disclosure. After receiving 1410 the combined mmW radar data, the target tracking module 1400 may be configured to segment 1420 the combined mmW radar data to detect a target. The segmenting can include partitioning points of the mmW radar data into groups corresponding to the target. For example, the combined mmW radar data may include a three-dimensional (3D) array of points (i.e., point cloud) with each point corresponding to a reflected signal from targets/objects in the combined door area. Segmentation can label certain points in the point cloud as corresponding to the target. Further, segmentation may include determining a boundary or boundaries in the combined mmW radar data that define the target. In other words, the target tracking module 1400 can be configured to detect a target in the combined door area. The detected targets may include people, animals, structures, and any other object/entity that can reflect signals back to the radar sensor. The mmW radar data corresponding to the detected targets may be further analyzed by the target tracking module to distinguish the targets further so that, for example, a person may be distinguished from an object (i.e., classified).

After a target is detected, the target tracking module may be further configured to determine 1430 physical characteristics of the target. For example, the target data may be analyzed to determine a shape of the target. The shape of the target may include a width of the target. Additionally, the floor maybe detected from the combined mmW radar data and the height of the target off the floor may be determined. Physical characteristics of the target may help to classify the target. For example, the target may be classified as living (e.g., a person) or as an inanimate (e.g., an object) based on its physical characteristics (e.g., height). These physical characteristics are provided as examples. It is recognized that other physical characteristics and/or information related to physical characteristics (e.g., inanimate/animate) may be determined for a target. Accordingly, this discussion is not intended to limit the present disclosure to these examples.

After a target is detected, the target tracking module may be further configured to determine 1440 a position of the target. The position of the target may be mapped to a position, or region, in the combined door area. Further the position of the target may include a mapping to a particular door area in the combined door area. For example, the position of the target may be designated as corresponding to a particular radar sensor. Further, the position of the target may include a mapping to a zone (e.g., activation zone 131, presence zone 132, etc.) within the combined door area. These positions are provided as examples. It is recognized that other positions and/or information related to a position (e.g., zone) may be determined for a target. Accordingly, this discussion is not intended to limit the present disclosure to these examples.

After a target is detected, the target tracking module may be further configured to determine 1450 a motion of the target. The motion of the target may include a velocity of the target, which includes a speed of the motion and a direction of the motion. When the motion of the target is combined with a position of the target, a path of the target may be determined. In some implementations, the mmW radar data may be analyzed to determine motion on a very small scale. For example, a breathing motion of a person may be detected and used to classify the target (e.g., as animate). These motions are provided as examples. It is recognized that other motions and/or information related to motion (e.g., breathing) may be determined for a target. Accordingly, this discussion is not intended to limit the present disclosure to these examples.

The target tracking module 1400 may be further configured to record (i.e., store) the physical characteristic information, the position information, and/or the motion information (i.e., target information) in a target database 1500. In a possible implementation, the target database 1500 may be queried based on the physical characteristic information, the position information, and/or the motion information to determine if the target is new or if the target is already recorded in the target database. For a target recognized as being in the database, the target tracking module 1400 may be configured to update target information stored for the target. The updates may also add or create new target information corresponding to the target. For example, previously obtained positions may be combined with presently obtained positions to compute a path of the target.

After the target database is updated for a detected target, the mmW radar data may be further analyzed for other targets and the process may be repeated. In other words, detecting targets and determining the corresponding target information may be carried out in series until all targets are recorded. In another possible implementation, segmenting the mmW radar data may include detecting multiple targets in one step and determining the target information for each of the multiple targets may be carried out in a parallel fashion.

FIG. 15 is a flowchart illustrating details of the door control module 1510, the door filter module 1530, and the door operation module 1550 according to a possible implementation of the present disclosure. The details will be described for a target, though in practice multiple targets may be processed simultaneously. The door control module 1510 (i.e., door control) and the door filter module 1530 (i.e., door filter) are configured to process the physical characteristic information, the position information, and/or the motion information of the target (e.g., from the target database 1500) in order to generate estimates (e.g., predictions) and identifications that can be used by a door operation module 1550 (i.e., door operation) to decide an action to be taken with the door. Additionally, a conclusion made by the door operation module may be based on door information 1560, such as a state of the door, a specification of the door, and/or a door model, which is a mathematical representation of the door based on its specifications.

As mentioned previously, the door control module 1510 can be configured to process target information to determine how, and when, the door should be controlled. For this, the door control module 1510 may include a crossing point estimator 1511. The crossing point estimator 1511 can be configured to estimate if the target will cross the threshold 115 of the door opening 103. When it is determined that the target will cross the threshold 115 of the door opening 103, the crossing point estimator 1511 can be further configured to estimate where, within the door opening 103, the target will cross the threshold 115 of the automatic door. These estimates may be based on the current position of the target and the target's direction of movement obtained from the target database 1500 (e.g., by query).

The door control module 1510 may further include a time of arrival estimator 1512. The time of arrival estimator 1512 can be configured to estimate when (i.e., a time) the target will arrive at the threshold of the automatic door. This estimate may be based on the position and velocity (i.e., speed, direction) of the target.

The door control module 1510 may further include a door opening estimator 1513. The door opening estimator 1513 can be configured to estimate how large a door opening is needed to accommodate the target. The estimated door opening can be less than a full (i.e., 100%) door opening. This estimate may be based on the size (e.g., width, height), and shape of the target. In some implementations, the door opening estimator 1513 can be configured to estimate where in a range of possible door openings the door opening should be located. This estimate may be based on the crossing point estimated for the target. The results of the estimators may be transmitted to the door operation module 1550. Additionally, the results of the estimators may be transmitted to the target database 1500 to further characterize the target.

As mentioned previously, the door filter module 1530 can be configured to process target information to identify certain conditions that could trigger certain door operations (or prevent certain door operations). These identifications can be used to filter the target. In other words, based on the identified conditions, the target may be considered or ignored for door control. The door filter module 1530 may include a parallel traffic identifier 1531 (i.e., parallel traffic filter). The parallel traffic identifier may determine motion of a target that is moving in a direction other than towards the door. For example, the parallel traffic identifier 1531 may be configured to determine that the target is moving in a direction that is substantially parallel with the door opening 103. The output of the parallel traffic identifier can be a signal having a signal level that indicates that parallel traffic has been identified (e.g., HIGH level) or that parallel traffic has not been identified (e.g., LOW level). The signal may trigger the door operation module 1550 to ignore a target that is not expected to interact with the door.

The door filter module 1530 may further include a nuisance identifier 1532 (i.e., nuisance filter). The nuisance identifier may be configured to determine motion of a target that can be ruled out as actionable movement for the automatic door. For example, the nuisance identifier may determine that the target is moving in a repetitive way, such as motion from a tree blowing in the wind. The output of the nuisance identifier 1532 can be a signal having a signal level that indicates that a nuisance has been identified (e.g., HIGH level) or that a nuisance has not been identified (e.g., LOW level). The signal may trigger the door operation module 1550 to ignore a nuisance target.

The door filter module 1530 may further include an inanimate identifier 1533. The inanimate identifier 1533 may determine that a target is inanimate (i.e., not living). For example, the inanimate identifier may determine that no breathing motion is associated with the target and/or that the height of the target is smaller than a threshold. The output of the inanimate identifier 1533 can be a signal having a signal level that indicates that the target present is inanimate (e.g., HIGH level) or that the target present is animate (e.g., LOW level). The signal may trigger the door operation module 1550 to ignore inanimate targets.

The door filter module 1530 may further include a gesture identifier 1534. The gesture identifier 1534 may identify motion as a particular gesture. For example, the gesture identifier may determine that a person is waving. The output of the gesture identifier 1534 can be a signal having a signal level that indicates that a motion of a target is a gesture (e.g., HIGH level) or that the motion of the target present is not a gesture (e.g., LOW level). Alternatively, the signal may indicate which gesture from a plurality of possible gestures is detected. The signal may trigger the door operation module 1550 to control the door (e.g., open the door) according to the gesture. The results of the identifiers (i.e., filters) may be transmitted to the door operation module 1550. Additionally, the results of the estimators may be transmitted to the target database 1500 to further characterize the target.

As mentioned previously, the door operation module 1550 can be configured to process one or more estimates generated by the door control module 1510, one or more identifications generated by the door filter module 1530, and door information 1560 to determine an action for the automatic door. The door information 1560 may include a state (e.g., percent opened) of the automatic door. The state of the door may be determined by a mathematical door model with inputs corresponding to the inputs of the door. The door operation module 1550 may include an action classifier 1551 that is configured to generate a door action based on the estimates, identifications, and the door information. In a possible implementation the action classifier 1551 is a machine learning model (e.g., neural network) configured to output a signal corresponding to an action of the door based on the estimates and identifications. The door operation module 1550 may further translate this signal corresponding to the action of the door into a signal, or signals, to control the door assembly.

FIG. 16 is a block diagram illustrating a possible implementation of the automatic door system. As shown, the DCU 210 can communicate with a memory (e.g., memory 211 included in the DCU) in which the target database 1500 is stored. The DCU 210 can be configured (e.g., by software) to access the target database 1500 to gather combined mmW radar data (i.e., radar data) about a target. The radar data can include three positions (i.e., P_xt, P_yt, P_zt) corresponding to three dimensions (3D) of a reference coordinate system. For the implementation described, the positions are registered so that P_zt=0 corresponds to a height of zero (i.e., at a floor level). The radar data may be processed to derive information about the movement of the target. For example, various velocities (i.e., speeds) of the target may be derived.

FIG. 17 is a diagram illustrating a mean speed (V_m) calculation according to a possible implementation of the present disclosure. The mean speed is computed as a change in position over a change in time (i.e., a rate of change in position). The mean speed is based on a position change vector ({right arrow over (dP)}). The position change vector ({right arrow over (dP)}) is in a direction along a line between a first position (P_xt-1, P_yt-1, 0) at a first time and a second position (P_xt, P_yt, 0) at a second time. The positions can be included in the registered radar data from a radar sensor 1720. As shown by equation 1710, the mean speed can then be computed as the magnitude of the position change vector ({right arrow over (|dP|)}) divided by the change in time between the first time and the second time (dt). In other words, the mean speed is the rate at which the target is moving towards a crossing point (i.e., shown in FIG. 17 as X) of the threshold of the automatic door.

A crossing point estimator 1511 may use the position change vector ({right arrow over (dP)}) to determine where the target will cross the threshold of the automatic door. As shown in FIG. 17, the position change vector ({right arrow over (dP)}) can be projected to a plane of the door in which the center of the door is at (C_x0, C_y0, 0) (i.e., shown in FIG. 17 as O) in order to estimate the crossing point (C_x, C_y, 0), which is also in the plane of the door. The center of the door may not be aligned with the crossing point.

A radar sensor 1720 may be position directly above the center of the door (i.e., may be aligned with the center of the door). The radar sensor can be configured to directly measure velocity (i.e., speed) of a target; however, this speed, which is known as the measured doppler speed (V_D), is in a direction towards the radar sensor. Accordingly, a mean speed (V_m) based on a calculation using positions measured by the radar sensor may be different from a measured doppler speed (V_D) based on a direct measurement of speed by the radar sensor.

FIG. 18 is a diagram illustrating a projected doppler speed (V_Dp) calculation according to an implementation of the present disclosure. As shown, a registered location (C_x0, C_y0, C_z0) of the radar sensor is above the door opening (i.e., above the center of the door). The radar sensor may be configured to measure a doppler speed (V_D) at which the target approaches (i.e., closes in on) the radar sensor 1720. In other words, the speed measured by the radar sensor may be a speed along a first vector component ({right arrow over (PS)}) towards the radar sensor 1720. The speed for computing a time of arrival is along a second vector component ({right arrow over (PC)}) towards the center of the door. As shown by equation 1810, the projected doppler speed (V_Dp) is a fraction of the measured doppler speed (V_D) in the direction towards the center of the door. In other words, the projected doppler speed is the speed component from the target location directed to the center of the door (at zero height).

Control of the automatic door may also be based on an estimated time of arrival (T_a) of the target to the door from its current position (P_T) and a time of arrival estimator 1512 can be configured to calculate a time of arrival using the mean speed (V_M) and/or the measured doppler speed (V_D). In other words, a time to arrive at the door can be computed as the time required to reach the crossing point (C