Ignition: An End-to-End Supervised Model for Training Simulated Self-Driving Vehicles
This work addresses autonomous driving in simulated environments, but it is incremental as it builds on existing ResNet architectures without major methodological breakthroughs.
The authors tackled training self-driving vehicles in simulation by introducing Ignition, an end-to-end neural network that processes front-facing images to output steering, throttle, and braking controls, demonstrating that latent road features can be automatically learned without explicit training.
We introduce Ignition: an end-to-end neural network architecture for training unconstrained self-driving vehicles in simulated environments. The model is a ResNet-18 variant, which is fed in images from the front of a simulated F1 car, and outputs optimal labels for steering, throttle, braking. Importantly, we never explicitly train the model to detect road features like the outline of a track or distance to other cars; instead, we illustrate that these latent features can be automatically encapsulated by the network.