An improved deep reinforcement learning-based approach for adaptive speed planning of unmanned vehicles that maintains optimal speed while maneuvering around obstacles.