Abstract:

As a new type of transportation， flying cars offer the advantages of rapid aerial flight and stable ground driving. However， most existing path planning algorithms rely on separate planning strategies for aerial and ground states， and followed by their combined superposition， which results in slow state transition and leads to frequent takeoffs and landings at switching points. To address these issues， in this paper a land-air integrated path planning method based on the state augmentation of sampling points is proposed. Considering the differences in land-air states and spatial environment， a classification strategy for path sampling points is established. Furthermore， land-air state information is incorporated to augment the traditional sampling points. This enables rapid and efficient transition between aerial and ground states during vehicle movement， while also reducing path planning computation time. The experimental results show that， compared to conventional combined path planning algorithms， the proposed method reduces travel distance by 22.3% and path planning time by 34.2%. Compared to stateless augmentation-based integrated planning algorithms， the proposed method achieves an 8.4% reduction in path length and an 11.1% decrease in motion cost.

Key words: flying cars, unmanned vehicle, path planning, sampling nodes state augmentation, trajectory optimization