V2VNet: Vehicle-to-Vehicle Communication for Joint Perception and Prediction
This addresses the challenge of limited sensor visibility in autonomous driving, though it appears incremental as it builds on existing V2V communication concepts with a specific compression method.
The paper tackles the problem of improving perception and motion forecasting for self-driving vehicles by using vehicle-to-vehicle communication to aggregate information from multiple viewpoints, enabling detection through occlusions and at long ranges with high accuracy while meeting bandwidth constraints.
In this paper, we explore the use of vehicle-to-vehicle (V2V) communication to improve the perception and motion forecasting performance of self-driving vehicles. By intelligently aggregating the information received from multiple nearby vehicles, we can observe the same scene from different viewpoints. This allows us to see through occlusions and detect actors at long range, where the observations are very sparse or non-existent. We also show that our approach of sending compressed deep feature map activations achieves high accuracy while satisfying communication bandwidth requirements.