New technologies for Autonomous Vehicles and Mobile Robots are presented, with an emphasis on Multi-sensors Embedded Perception, Situation Awareness, Collision Risk Assessment, and Decision-making for safe navigation in Dynamic Human Environments. It is shown that Bayesian approaches are mandatory to develop such technologies and to obtain the required robustness in presence of uncertainty and of complex dynamic situations. It is also shown how the research results have been used to tackle some practical industrial problems.