Consumer-Electronics-Oriented Safe Interaction-Aware Motion Planning for Ramp Driving with Perception Uncertainty Quantification

Autonomous ramp driving is challenging due to complex interactions between autonomous vehicles (AVs) and human-driven vehicles (HDVs) with diverse driving styles, as well as strict geometric constraints imposed by ramp terminations. For consumer electronic vehicles (CEVs), these challenges are further amplified by limited computational resources and real-time decision requirements. This paper proposes a consum-erelectronics-oriented cooperative motion planning framework for autonomous ramp driving that integrates driving style awareness, perception uncertainty quantification, and interaction-aware decision-making. Heterogeneous HDV behaviors are modeled using data-driven driving style classification, while prediction uncertainty is captured via error ellipses. A game-theoretic formulation is employed to select lane-changing time points, which are then translated into executable trajectories through lightweight planning and control modules. Simulation results based on the NGSIM dataset and diverse ramp traffic scenarios demonstrate collision-free operation with improved merge efficiency and velocity smoothness under varying traffic densities. Furthermore, a dedicated consumer electronics platform evaluation shows that the proposed method achieves sub-millisecond runtime, minimal memory usage, and low energy consumption, outperforming several recent 2025 state-of-the-art planners, including mean-field-game-based MPC, Monte Carlo decision processes, and risk-field-based game-theoretic approaches. These results indicate that the proposed framework is well suited for safe and efficient ramp driving on consumer-grade autonomous vehicle platforms.