Dynamic lane-changing trajectory planning for autonomous vehicles: A novel approach using Gravitational Search Algorithm with Blend Crossover

With the rapid developments in autonomous driving technology, the safe and efficient operation of autonomous vehicles is ensured using a highly potent critical component called dynamic lane-changing trajectory planning. Likewise, we present a novel approach using the Gravitational Search Algorithm with Blend Crossover (GSA-Blend) to tackle the challenges related to real-time trajectory optimization in dynamic and complex driving scenarios. Both the principles of gravitational search and blend crossover techniques are leveraged in the proposed GSA-Blend algorithm to dynamically plan lane-changing trajectories. Using GSA-Blend, the comprehensive cost function is optimized to enhance passenger comfort, safety, and efficiency, aiming to achieve a balanced trade-off among these key aspects of autonomous driving. Furthermore, a new blend crossover operator introduced, which has tailored lane changing trajectory planning to improve the algorithm's performance under complex scenarios. The effectiveness of GSA-Blend is validated through conducting extensive simulations in diverse driving scenarios, including hazardous, complex and simple conditions. The algorithms performance was evaluated against existing methods, considering lane-changing efficiency, safety, and comfort metrics. The obtained results have clearly shown that GSA-Blend consistently outperformed conventional methods in terms of its adaptability and robustness across various real-world scenarios. This research advances autonomous driving through introducing an innovative trajectory planning solution that enhances real-time optimization, ensuring safer and efficient lane changes. Ultimately, reliable and safer transportation systems can be achieved with the improvement of autonomous vehicle abilities using the promising potentials offered by the GSA-Blend algorithm.