Hyundai Mobis has developed a new passenger detection system that uses radar technology to monitor people riding in the back seat. The company is planning to pitch the Rear Occupant Alert (ROA) system to other auto manufacturers, although the solution has not yet entered the mass production phase.

The ROA system was designed to prevent people from leaving infants unattended in the back seat of a vehicle. It will notify the driver if it notices that someone is still in the back seat after that driver gets out of the car, either by making the sound the car typically makes when the door is closing or by sending an alert to the driver’s smartphone or another device.  

The system is sensitive enough to distinguish infants from other riders like adults and pets. Hyundai Mobis argues that it is superior to contemporary rider alert systems, which typically utilize an ultrasonic sensor or a weight sensor in a child’s car seat. The new ROA system replaces those sensors with a radar sensor that can detect the micromovements of passengers, such as a rising chest that indicates that someone is breathing.

“As the radar can penetrate clothes and measure various biosignals, it can detect the presence of passengers in rear seats more accurately,” said Jang Jae-ho, the Director of the EE Research Center for Hyundai Mobis. “It overcame the weakness of the camera sensor, which cannot recognize babies covered in blankets.”

The software algorithm is based on radar sensor technology that was being developed for autonomous vehicles. It has secured electromagnetic reliability, and will function as intended near railroad tracks and other high-voltage areas. Hyundai Mobis also indicated that upcoming improvements will allow the system to measure heartbeats and other biometric functions.

The ROA system will supplement Hyundai Mobis’ existing Driver State Warning system, which tracks the driver to make sure they are paying attention to the road. The DSW system debuted in July, several months after Hyundai Mobis made a $5 million investment in the AI and computer vision specialist Deep Glint.