Multimodal Classification Network Guided Trajectory Planning for Four-Wheel Independent Steering Autonomous Parking Considering Obstacle Attributes

TL;DR

<3-5 sentence high-level summary> The paper tackles autonomous parking for four-wheel independent steering (4WIS) vehicles in narrow, obstacle-dense spaces by integrating a multimodal classification network (MCN) to gauge scene difficulty, guided point generation to decompose hard tasks, and a 4WIS-specific hybrid A* search with obstacle-aware expansion. A hierarchical obstacle handling strategy prioritizes crossing, then driving over, and finally avoidance, complemented by probabilistic risk-field driving corridors to handle dynamic obstacles. The initial path from the 4WIS hybrid A* warms-starts an optimal control problem that optimizes safe, smooth trajectories, with logical constraints for drive-over obstacles. Experimental results show superior success rates, shorter paths, faster computation, and improved safety across varying obstacle densities compared with baseline methods, highlighting the framework’s practical value for constrained autonomous parking.

Abstract

Four-wheel Independent Steering (4WIS) vehicles have attracted increasing attention for their superior maneuverability. Human drivers typically choose to cross or drive over the low-profile obstacles (e.g., plastic bags) to efficiently navigate through narrow spaces, while existing planners neglect obstacle attributes, causing inefficiency or path-finding failures. To address this, we propose a trajectory planning framework integrating the 4WIS hybrid A* and Optimal Control Problem (OCP), in which the hybrid A* provides an initial path to enhance the OCP solution. Specifically, a multimodal classification network is introduced to assess scene complexity (hard/easy task) by fusing image and vehicle state data. For hard tasks, guided points are set to decompose complex tasks into local subtasks, improving the search efficiency of 4WIS hybrid A*. The multiple steering modes of 4WIS vehicles (Ackermann, diagonal, and zero-turn) are also incorporated into node expansion and heuristic designs. Moreover, a hierarchical obstacle handling strategy is designed to guide the node expansion considering obstacle attributes, i.e., 'non-traversable', 'crossable', and 'drive-over' obstacles. It allows crossing or driving over obstacles instead of the 'avoid-only' strategy, greatly enhancing success rates of pathfinding. We also design a logical constraint for the 'drive-over' obstacle by limiting its velocity to ensure safety. Furthermore, to address dynamic obstacles with motion uncertainty, we introduce a probabilistic risk field model, constructing risk-aware driving corridors that serve as linear collision constraints in OCP. Experimental results demonstrate the proposed framework's effectiveness in generating safe, efficient, and smooth trajectories for 4WIS vehicles, especially in constrained environments.

Figures (12)

• Figure 1: Trajectory planning of the front-wheel steering vehicle and the 4WIS vehicle. In contrast, the 4WIS vehicle successfully navigates the narrow space by leveraging multiple motion capabilities.
• Figure 2: Overview of the framework. We use image and vehicle state as inputs, and employ a multimodal classification network to assess the complexity of the planning task. For hard tasks, a guided point generation strategy is activated to facilitate the 4WIS hybrid A* path planning, which accommodates multiple motion modes and incorporates hierarchical obstacle handling strategies. The 4WIS hybrid A* initial path is a warm start for solving the OCP.
• Figure 3: 4WIS motion modes. (a) Ackermann steering mode; (b) Diagonal movement mode; (c) Zero-turn rotation mode.
• Figure 4: Adaptive step size strategy for driving corridor generation. In the ($i-1$) th iteration step, the step size is 2$\Delta s$ in the rightward and upward direction. A collision encountered in the upward direction triggers a reduction in step size, which continues until a collision-free expansion is achieved. Conversely, in the rightward direction, the step size increases iteratively until a collision is detected, after which it is reduced to enable further exploration. The generated driving corridor is shown in (d) after successful expansion.
• Figure 5: Guided points setting. (a) Visible points. (b) Gear shifting points.