US Patent 11529966 Risk prediction on a peer-to-peer network

A method for controlling a driving behavior performed by a first agent includes navigating, by the first agent, according to a trajectory and a velocity. The method also includes receiving, from a second agent, a risk identification message identifying a third agent as a potential risk based on the third agent performing a behavior associated with a probability that is less than a threshold. The method further includes autonomously engaging a defensive driving mode in response to receiving the risk identification message. The method still further includes adjusting one or both of the trajectory or the velocity in response to autonomously engaging the defensive driving mode.