A Global Path Planner for Safe Navigation of Autonomous Vehicles in Uncertain Environments.

Area of Science:

• Robotics and Artificial Intelligence
• Autonomous Systems Engineering

Background:

• Autonomous vehicle (AV) adoption is hindered by navigation challenges under sensor uncertainty.
• No single sensor solution is perfect for all driving conditions, necessitating robust planning.

Purpose of the Study:

• To propose a global safe path planner that analyzes sensor uncertainty for optimal path determination.
• To improve the safety and reliability of AV navigation in dynamic environments.

Main Methods:

• Developed a two-component path planner: sensor analytics and path finder.
• Sensor analytics component quantifies positioning and navigation performance based on combined sensor uncertainties.
• Path finder component assigns safety weights to road segments using sensor performance data.

Main Results:

• Simulations demonstrated the effectiveness of the proposed safe path planner.
• The planner generated paths that significantly improved navigation safety compared to conventional methods.
• The system proved effective in complex dynamic driving environments.

Conclusions:

• The proposed global safe path planner effectively addresses sensor uncertainty in AV navigation.
• This approach offers a significant advancement in ensuring safe autonomous driving.
• The findings support the expedited adoption of autonomous vehicles.