Locally Weighted Regression Pseudo-Rehearsal for Online Learning of Vehicle Dynamics
This addresses the challenge of preventing catastrophic forgetting in online learning for autonomous vehicle control, which is crucial for safe and reliable operation, though it appears incremental as it builds on existing pseudo-rehearsal and regression techniques.
The paper tackles the problem of online adaptation of a neural network for vehicle dynamics, which faces catastrophic forgetting due to non-stationary distributions, by proposing a novel method combining pseudo-rehearsal with locally weighted projection regression, demonstrating effectiveness in simulation and on a real-world dataset from a 1/5 scale autonomous vehicle.
We consider the problem of online adaptation of a neural network designed to represent vehicle dynamics. The neural network model is intended to be used by an MPC control law to autonomously control the vehicle. This problem is challenging because both the input and target distributions are non-stationary, and naive approaches to online adaptation result in catastrophic forgetting, which can in turn lead to controller failures. We present a novel online learning method, which combines the pseudo-rehearsal method with locally weighted projection regression. We demonstrate the effectiveness of the resulting Locally Weighted Projection Regression Pseudo-Rehearsal (LW-PR$^2$) method in simulation and on a large real world dataset collected with a 1/5 scale autonomous vehicle.