Sensor Fusion and autonomous vehicles
The Edge AI and Vision Alliance held a webinar on the topic of "Sensor Fusion and Autonomous Vehicles" which showed the importance radar technology plays in autonomous vehicles.
Here is a summary of the key (radar related) points.:
- 77-79GHz is the most common frequency for vehicle radar. Vehicle radar is used for Adaptive Cruise Control, Automatic Emergency Braking, Blind Spot Detection, Child Presence Detection, Forward Collision Warning, Lane Change Assist, Park Assist, Rear Cross Traffic Alert, Rear Collision Alert, Traffic Jam Assist, Vehicle Exit Assist, and Vulnerable Road Use.
- Radars are starting to get elevation capability and are getting higher resolution
- AI and image processing of radar data is going to be one of the next big trends.
- The 2019 total market size for Vehicle Radar was around $5B and this is set to double in the next 5 years.
- Lidar has a very small slice of ADAS sensor market compared to cameras, and radar.
- One of the key challenges with sensor fusion is reconciling the different fields of view from each sensor (as they are positioned on different parts of a vehicle). This is know as the parallax effect.
- Sensor processing is highly distributed (each sub-system does its own processing) at the moment but the trend is to centralize the processing in one place.
- Vendors are coming out with specialist hardware for sensor fusion tasks.