Aiming at the Connected Autonomous Vehicles (CAVs) crossing the unsignalized intersection problem, a hybrid coordinated optimization method is studied in this paper. The proposed approach consists of a Multi-Risk Management of Cooperative Optimization approach based on the Predicted Inter-Distance Profile (MRMCO-PIDP), and Epsilon Probability Collective algorithm (Epsilon-PC) to make CAVs apply cooperative and adaptive velocity planning to navigate safely and quickly in unsignalized intersections. It is the first use of PIDP for multi-risk assessment and management for CAVs. According to the computation of PIDP metric [1] regarding other vehicles with risk of collision and its controlled minimum (mPIDP), CAVs can find the most effective speed profile for collision avoidance in an unsignalized intersection. Several random scenarios are performed in simulation to demonstrate the reliability of the proposed approach.