Fast Incremental Learning for Off-Road Robot Navigation
This addresses the data efficiency challenge for off-road robot navigation, though it appears incremental as it builds on existing transfer learning approaches.
The paper tackles the problem of requiring large training datasets for autonomous robot navigation by developing a system that leverages ImageNet's pre-trained models and adapts quickly with minimal environment-specific data.
A promising approach to autonomous driving is machine learning. In such systems, training datasets are created that capture the sensory input to a vehicle as well as the desired response. A disadvantage of using a learned navigation system is that the learning process itself may require a huge number of training examples and a large amount of computing. To avoid the need to collect a large training set of driving examples, we describe a system that takes advantage of the huge number of training examples provided by ImageNet, but is able to adapt quickly using a small training set for the specific driving environment.