Predicting Future Lane Changes of Other Highway Vehicles using RNN-based Deep Models
share
In the event of sensor failure, it is necessary for autonomous vehicles to safely execute emergency maneuvers while avoiding other vehicles on the road. In order to accomplish this, the sensor-failed vehicle must predict the future semantic behaviors of other drivers, such as lane changes, as well as their future trajectories given a small window of past sensor observations. We address the first issue of semantic behavior prediction in this paper, by introducing a prediction framework that leverages the power of recurrent neural networks (RNNs) and graphical models. Our prediction goal is to predict the future categorical driving intent, for lane changes, of neighboring vehicles up to three seconds into the future given as little as a one-second window of past LIDAR, GPS, inertial, and map data. We collect real-world data containing over 500,000 samples of highway driving using an autonomous Toyota vehicle. We propose a pair of models that leverage RNNs: first, a monolithic RNN model that tries to directly map inputs to future behavior through a long-short-term-memory network. Second, we propose a composite RNN model by adopting the methodology of Structural Recurrent Neural Networks (RNNs) to learn factor functions and take advantage of both the high-level structure of graphical models and the sequence modeling power of RNNs, which we expect to afford more transparent modeling and activity than the monolithic RNN. To demonstrate our approach, we validate our models using authentic interstate highway driving to predict the future lane change maneuvers of other vehicles neighboring our autonomous vehicle. We find that both RNN models outperform baselines, and they outperform each other in certain conditions.
READ FULL TEXT