Researchers at the Technical University of Munich (TUM) have developed a new early warning system for vehicles that uses artificial intelligence to learn from thousands of real traffic situations.
A study of the system, developed with Prof. Eckehard Steinbach, who holds the Chair of Media Technology and is a member of the Board of Directors of the Munich School of Robotics and Machine Intelligence (MSRM) at TUM, with the BMW Group showed it can warn drivers up to seven seconds in advance against potentially critical situations that the cars cannot handle alone with over 85% accuracy.
The central storage of the data makes it possible for every vehicle to learn from all of the data recorded across the entire fleet, which also has implications for 5G links to cars.
he introspective failure prediction approach is based on a recurrent neural network (RNN) that uses the disengagements triggered either by automatic safety measures or by human intervention as training data to learn to predict future failures. The system then learns introspectively from its own previous mistakes.
The system uses two sensor types. An image-based model learns to detect generally challenging situations such as crowded intersections accurately multiple seconds in advance. A state data based model allows to detect fast changes immediately before a failure, such as sudden braking or swerving.
The technology uses sensors and cameras to capture surrounding conditions and records status data for the vehicle such as the steering wheel angle, road conditions, weather, visibility and speed and the RNN framework learns to recognize patterns with the data. If the system spots a pattern in a new driving situation that the control system was unable to handle in the past, the driver will be warned in advance of a possible critical situation.
"To make vehicles more autonomous, many existing methods study what the cars now understand about traffic and then try to improve the models used by them. The big advantage of our technology: we completely ignore what the car thinks. Instead we limit ourselves to the data based on what actually happens and look for patterns," said Steinbach. "In this way, the AI discovers potentially critical situations that models may not be capable of recognizing, or have yet to discover. Our system therefore offers a safety function that knows when and where the cars have weaknesses."
The outcome of the individual models is fused by averaging the individual failure probabilities. The team of researchers tested the technology with the BMW Group and its autonomous development vehicles on public roads and analyzed around 2500 situations where the driver had to intervene. The study showed that the AI is already capable of predicting potentially critical situations with better than 85 percent accuracy - up to seven seconds before they occur.
For the technology to function, large quantities of data are needed. However the large number of development vehicles on the road means the data was practically generated by itself, says Christopher Kuhn, one of the authors of the study. "Every time a potentially critical situation comes up on a test drive, we end up with a new training example," he said.
- Self-healing software for driverless cars
- Extensible DSP for autonomous driving
- Microsoft, Volkswagen join forces for automated driving
- Autonomous racing challenge to use RTI software
Other articles on eeNews Europe