TI - Detecting occupancy in moving vehicles using mmWave sensors

Aguilera

TI - Detecting occupancy in moving vehicles using mmWave sensors [Copy link]

Automotive designers have successfully integrated millimeter wave (mmWave) sensors into several automotive in-cabin applications. One of these applications is the ability to detect occupancy in a vehicle, regardless of movement, across a wide range of lighting conditions and sensor placements. This can help automotive systems detect the location of children or people left unattended in the vehicle for climate control purposes. Azcom Technology demonstrated how the AWR1642 mmWave sensor combined with Azcom’s proprietary algorithms can reliably identify seat occupancy. We drove the vehicle at different speeds and in different environmental (city, highway) and cabin (lighting, temperature) conditions, and analyzed different seat configurations. In our demonstration, the sensor will be suspended from the sunroof, toward the back seat (as shown in Figure 1), although in a final installation it will more likely be placed inside a seatback, around a rearview mirror, or inside the roof. Because mmWave can sense a variety of materials, including those that make up the vehicle, the sensing performance does not change when installed inside a seat or roof. All processing, including Azcom Technology enhancements, runs on the sensor, while a graphical user interface on the host computer helps visualize the results. Figure 1: mmWave sensor mounted on vehicle sunroofThe main challenge for this use case is to achieve sufficient detection robustness when the engine is on and the car is moving. The combination of these two events introduces a set of interruptions in the signal from several vibration modes that are not present in a static setting. For this reason, a new algorithm was designed. This algorithm is less sensitive to vibrations from the road and is able to detect all possible combinations of occupancy conditions. These enhancements were applied and validated in addition to the occupancy detection reference design. Figures 2, 3, 4, and 5 are some snapshots from a sample drive and an illustration of occupancy detection. In Figure 2, the algorithm detects no faults when the rear seats are empty while driving in the city. Statistics are calculated at the rate of processed frames: 6 fps in this use case. In a real product, a less frequent secondary decision tool would make the detection more robust. Figure 2: Statistics for unoccupied rear seats Failure rate for zone 1 0.00% Failure rate for zone 2 0.00% Statistics calculated via GUI Counters Use case 1: Both rear seats unoccupied In Figure 3, the algorithm successfully detects that someone is sitting in zone 1, as shown in the red box. Zone 1 Zone 2 Statistics Zone 1 Failure Rate 0.00% Zone 2 Failure Rate 0.00% Statistics calculated via GUI counter Use Case 1: No Occupants in Both Rear Seats Figure 3: Occupants Detected in Rear Seats Figure 4 focuses on the accuracy of the algorithm when using different car cabs. By testing two different car models, we demonstrate reliable occupancy detection while driving at different speeds. Figure 5 shows the design extended to two rows of four seats. Although this scenario is more complex and challenging, after special adjustments and optimizations, the algorithm performs just as well as in the single-row setting. Figure 5: Four-seat configuration 3 zones 4 zones 2 zones 1 zone Four zones Two people sitting in the statistics Failure rate of zone 1 0.00% Failure rate of zone 2 0.00% Failure rate of zone 3 0.00% Failure rate of zone 4 0.00% With expertise in occupancy detection applications and deep expertise in TI platforms, signal processing, RF and embedded system design and development, Azcom Technology offers a range of value-added R&D services to help you build mmWave-enabled products and significantly reduce your time to market. To better understand Azcom's portfolio of design and development services for mmWave sensor product development and support.