RAW SENSOR DATA SHARING FOR ENHANCED FLEET-WIDE ENVIRONMENTAL AWARENESS AND SAFETY

摘要

In one embodiment, a first vehicle makes, using a first machine learning-based model, a determination that raw image data captured by the first vehicle does not indicate the presence of a driving hazard. The first vehicle sends a hazard assessment request to a second vehicle in communication proximity of the first vehicle that includes the raw image data. The first vehicle receives a hazard detection response from the second vehicle that indicates a driving hazard identified by the second vehicle by applying a second machine learning-based model to the raw image data included in the hazard assessment request. The first vehicle adjusts a path of travel of the first vehicle to avoid the driving hazard identified by the second vehicle.