• Martin Booth

Safety, AI-friendly, and Upgradeable In-Cabin Technology

The European New Car Assessment Programme (Euro NCAP) is arguably the most active in promoting Artificial Intelligence (AI) based in-cabin automotive safety technologies and the legislation to back themi. Euro NCAP developed a roadmapii highlighting that more than 90 percent of road accidents are caused by either a violation – speeding, DUI, human error – in which the driver state (inattentiveness, fatigue, distraction) and inexperience play an essential role.

The need for help with driver safety is profound. One study of emergency room injuries from the year before and following the introduction of seatbelt legislation found that the number of those who escaped injury during auto collisions improved by 40 percentiii. According to CBS News, rear automatic braking reduced backup accidents in cars by 62 percent; adding on rearview cameras (required as of May 2018) boosted that number to 78 percentiv.

A range of new, AI-fueled in-cabin technologies stands poised to assist driver safety needs, including:

  • Adaptive Cruise Control (ACC) is a blending of cameras, radar, and various lasers that enable vehicles to intelligently maintain a proper distance and keep pace with cars ahead when the driver fails to maintain a safe distance manually. 
  • Driver Monitoring Systems weighs a variety of factors (e.g., road conditions, steering response patterns) assessing whether a driver is dozing off at the wheel. It then alerts accordingly utilizing camera-based input systems that can issue driver warnings, especially when monitoring in-cabin behaviors such as interacting with a cell phone.
  • Child Presence Detection (CPD) helps prevent heat stroke and death from overheating inside a sealed car by detecting for an unattended occupant. Some systems continue to monitor the rear seat for motion after the vehicle is parked and alerts the car owner or emergency services accordingly.

Because automakers and buyers alike crave the safety ratings that come from incorporating such abilities, these features and many more like them have been slowly making their way into vehicles, especially high-end models, over the past decade. However, this is only the tip of the safety iceberg.

AI powers an increasing number of next-gen safety and autonomous vehicle features. This proliferation requires a new type of computing – architecture and software platforms built for the demands and complexities of AI workloads. Blaize automotive-grade solutions are ideally suited to meet AI workloads demands of in-cabin and driver safety features. Fully programmable Blaize Graph Streaming Processor™ (GSP) devices enable flexible and upgradeable in-cabin & driver assist systems with features that can be rolled out over the air and updated automatically. The unique low-latency highly parallelized GSP architecture delivers up to 10-100X systems efficiency compared to legacy generation Graphics Processing Unit (GPU) based solutions.

For applications ranging across advanced driver-assistance systems, security, and gesture-based driver commands, Thinci GSP architecture is ready to provide OEMs and Tier-1 suppliers with a streamlined, affordable, future-ready path into tomorrow’s most demanded and required in-cabin & driver-assist AI services.