To meet growing consumer demand for more comfortable and feature-rich driving experiences, OEMs face increasing challenges in expanding in-cabin safety systems. These challenges stem from the need to comply with evolving regulatory requirements while minimizing design complexity and costs. Upcoming changes to safety standards, such as those from the European New Car Assessment Programme (Euro NCAP), will alter how new vehicles are evaluated, encouraging OEMs to integrate advanced sensing capabilities into their designs.

 

AWRL6844

 

Traditionally, expanding in-cabin sensing applications for features such as occupant monitoring, child presence detection, and intrusion detection required multiple independent sensors. However, the latest innovations in radar sensor SoCs now enable support for multiple functionalities within a single device. By leveraging advanced edge AI capabilities and trained algorithms for local data processing, these sensors help automotive system designers tackle design complexity and significantly reduce system costs.

 

Challenges in In-Cabin Radar Design

OEMs are currently focusing on three critical in-cabin sensing applications: occupant monitoring for seat belt reminders, child presence detection, and intrusion detection. These features are essential to ensuring passenger safety throughout the driving experience. However, implementing these features presents several design challenges:

 

Occupant Monitoring for Seat Belt Reminders: Existing systems rely on weight grid sensors embedded in seats, which require calibration for each individual seat. While these systems were historically limited to front-row seats, OEMs are now integrating sensors in rear seats to enhance safety and comply with regulatory requirements. This shift significantly increases the number of sensors in the vehicle, as well as the associated wiring and calibration time. Moreover, weight grid sensors cannot differentiate between living objects and inanimate items (e.g., a backpack), leading to false occupancy alerts that may impact the user experience.

Child Presence Detection: The upcoming 2025 Euro NCAP standards mandate direct child presence detection for safety ratings in parked vehicles. OEMs can integrate additional sensors to meet these requirements, but achieving the necessary performance often requires multiple ultra-wideband (UWB) sensors. Furthermore, systems with insufficient resolution may struggle to distinguish between children and adults, posing an additional design challenge.

Intrusion Detection: Intrusion detection systems are becoming more common in premium vehicles. These systems, certified by organizations like Thatcham Research, often use ultrasonic sensors to detect intrusions. However, simple systems are prone to false triggers caused by non-intrusive actions, such as someone walking near the car or vehicle movement due to external activities.

 

Addressing Challenges with Edge AI

To meet the increasing performance demands of in-cabin sensing applications, OEMs are adopting innovative technologies that lower costs and simplify design. The AWRL6844 60GHz mmWave radar sensor addresses these challenges within a single device, reducing system costs by up to $20 per unit. A comparative analysis of traditional sensor distributions versus a simplified single-sensor approach using the AWRL6844 highlights the efficiency gains achieved by this solution.

 

The AWRL6844 features 16 virtual channels that enable high spatial resolution for detecting and localizing occupants in the vehicle, even while in motion. AI processing of high-resolution data allows the radar to distinguish between living objects and inanimate items, reducing false detections. Intelligent clustering algorithms running on an integrated digital signal processor filter out noise caused by vehicle motion, enabling highly accurate detection of occupants. This capability has been demonstrated by distinguishing between passengers in rear seats and a pile of water bottles.

 

The AWRL6844 also features a wide field of view, which enhances its ability to detect occupants in typically blind areas, such as footwells and rear-facing child seats. The sensor’s hybrid processing approach combines traditional radar techniques with real-time machine learning models to classify adults and children effectively. This method enables faster tuning and model updates to accommodate new test cases or requirements, helping OEMs reduce deployment times. TI’s physical-information neural network ensures classification accuracy exceeding 90%, supporting reliable decision-making in complex scenarios.

 

Enhanced Intrusion Detection and Noise Reduction

For intrusion detection, the AWRL6844 integrates low-power modes and machine learning capabilities to enhance performance while minimizing battery drain. Consuming less than 50mW, the sensor can process up to 10 intrusion events per second without depleting the vehicle’s battery, making it particularly valuable in battery-powered electric vehicles. Additionally, its on-chip accelerators minimize interference with other device cores, reducing detection latency and maintaining high precision. The AWRL6844’s ability to filter environmental noise further reduces false alarms caused by vehicle movement or external factors, improving the reliability of intrusion detection systems.

 

Conclusion

For OEMs, balancing stringent safety requirements with cost constraints remains a significant challenge. The AWRL6844 radar sensor offers a scalable solution, supporting both low-power and high-performance applications without the complexity of integrating multiple independent technologies. Its superior detection, localization, and classification capabilities, combined with improved false detection performance, enable a seamless and enhanced user experience. By adopting innovative solutions like the AWRL6844, OEMs can design safer, smarter vehicles that meet the demands of modern regulations and consumer expectations.

AWRL6844 AWRL6844

TI

3~7 Days

In Stock:Join the Waitlist


Stay updated with Lisleapex by signing up for the newsletter

Insights submitbox