Path planning is an essential research topic in the navigation of mobile robots. Currently, rapidly-exploring random tree star (RRT*) and its variants are known for their probabilistic completeness and asymptotic optimality, making them effective in finding solutions for many path planning problems. However, slow convergence rate of the RRT* limits its practical efficiency. To address this problem, this paper proposed an enhanced RRT* algorithm by refining the extension process of the exploring tree. This enhancement aims to guide the tree approaching to obstacles (GAO) while exploring toward the target point. First, GAO-RRT* employed a dual-weighted sample strategy instead of random sample to guide search direction of the exploring tree. Second, a variable step size extension strategy was adopted to increase the efficiency of node generation, balancing searching time and path safety in regions with different obstacles densities. Third, growth status of new nodes was monitored in real-time, and a reverse growth strategy was proposed to guide the exploring tree to escape local optima. In addition, parent node creation procedure for new nodes was used to produce a better initial path. Finally, the proposed GAO-RRT* was compared with three state of the art algorithms on 16 different instances of four representative environments. Compared to RRT*, Quick-RRT* (Q-RRT*), and Fast-RRT* (F-RRT*), the results showed that (1) the average path cost of initial solutions obtained by GAO-RRT* decreased by 38.32%, 29.69%, and 20.44%, respectively; and (2) the average convergence time of solution obtained by GAO-RRT* to suboptimal (1.05* $C_{best}$) was reduced by 71.22%, 69.69%, and 58.37%, respectively. Simulation results indicated that GAO-RRT* outperforms the compared algorithms in terms of path cost and convergence speed.

Effective path planning is crucial for mobile robots to quickly reach rescue destination and complete rescue tasks in a post-disaster scenario. In this study, we investigated the post-disaster rescue path planning problem and modeled this problem as a variant of the travel salesman problem (TSP) with life-strength constraints. To address this problem, we proposed an improved iterated greedy (IIG) algorithm. First, a push-forward insertion heuristic (PFIH) strategy was employed to generate a high-quality initial solution. Second, a greedy-based insertion strategy was designed and used in the destruction-construction stage to increase the algorithm’s exploration ability. Furthermore, three problem-specific swap operators were developed to improve the algorithm’s exploitation ability. Additionally, an improved simulated annealing (SA) strategy was used as an acceptance criterion to effectively prevent the algorithm from falling into local optima. To verify the effectiveness of the proposed algorithm, the Solomon dataset was extended to generate 27 instances for simulation. Finally, the proposed IIG was compared with five state-of-the-art algorithms. The parameter analysis was conducted using the design of experiments (DOE) Taguchi method, and the effectiveness analysis of each component has been verified one by one. Simulation results indicate that IIG outperforms the compared algorithms in terms of the number of rescue survivors and convergence speed, proving the effectiveness of the proposed algorithm.