This paper proposes a two-layer trajectory optimization method for the autonomous ground vehicle (AGV). This twolayer strategy includes an efficient path planning layer and a fast trajectory planning ...layer. In the first layer, a novel target area adaptive rapidly exploring random tree algorithm (TAA-RRT*) is proposed to search the shortest path. This layer mainly includes a preprocessing and a sampling planning process. In the preprocessing process, the generalized voronoi diagram (GVD) is used to construct the environment information and find the initial path. Then, the sampled target area (TA) is constructed based on this initial path to provide non-uniform sampling. In the sampling planning process, the improved adaptive RRT* algorithm is used to carry out sampling planning in the TA, and the direct connection strategy (DCS) is combined to quickly locate the optimal solution. In the trajectory planning layer, combined with the constraints of the unmanned vehicle and the path constraints obtained in the first layer, the speed planning and the trajectory optimization are addressed by solving the optimal control problem (OCP). After performing a large number of experiments, the feasibility and effectiveness of the proposed method is verified.
This article presents a reliable and robust rapidly exploring random tree (R2-RRT*) algorithm to tackle challenges in mission planning of off-road autonomous ground vehicles (AGVs) under uncertain ...terrain environment. Two types of mobility reliability metrics, namely state mobility reliability (SMR) and mission mobility reliability (MMR), are first defined to quantify the mobility reliability of an AGV and to incorporate mobility reliability into mission planning. SMR measures the probability that a vehicle can pass through a specific location on a map of interest, whereas MMR quantifies mobility reliability of a mission path with the consideration of dependence of soil properties and slope over space. Based on the defined SMR and MMR metrics, two reliability-based robust mission planning models are developed to identify optimal paths that have robust travel time and satisfy specific reliability requirements. Moreover, a reliability-based path smoothing algorithm is developed to address the suboptimality of R2-RRT*. Results of a case study demonstrate the efficacy of the proposed models and algorithms. Note to Practitioners -This article was motivated to explicitly account for the uncertain terrain environment in mission planning of off-road autonomous ground vehicles (AGVs). Existing approaches, e.g., RRT and its variants, in general, oversimplify the uncertainty sources and overlook the reliability of vehicle mobility. These simplifications could lead to failure (i.e., immobility) of off-road AGVs on the obtained paths. This article suggests a reliability-based mission planning model by incorporating the proposed SMR and MMR metrics into mission planning considering the spatial-dependent uncertainty sources. To identify a reliable and robust path, we extend RRT* to R2-RRT* that achieves a tradeoff between mission cost and reliability. The proposed reliability-based mission planning model, however, is not limited to RRT* and can also be integrated with other path planning algorithms. It can be also applied to other unmanned vehicles and robots, such as the motion planning of unmanned aerial vehicles (UAVs) in adverse weather conditions.
•Propose a sampling-based asymptotically optimal path planning algorithm.•The proposed algorithm guarantees a fast convergence rate.•Theoretical proof of asymptotic optimality and fast convergence ...rate is given.
During the last decade, sampling-based algorithms for path planning have gained considerable attention. The RRT*, a variant of RRT (rapidly-exploring random trees), is of particular concern to researchers due to its asymptotic optimality. However, the limits of the slow convergence rate of RRT* makes it inefficient for applications. For the purposes of overcoming these limitations, this paper proposes a novel algorithm, PQ-RRT*, which combines the strengths of P-RRT* (potential functions based RRT*) and Quick-RRT*. PQ-RRT* guarantees a fast convergence to an optimal solution and generates a better initial solution. The asymptotic optimality and fast convergence of the proposed algorithm are proved in this paper. Comparisons of PQ-RRT* with P-RRT* and Quick-RRT* in four benchmarks verify the effectiveness of the proposed algorithm.
In this article, a novel manipulability-based optimal rapidly exploring random tree (RRT*) path planning strategy is proposed for industrial robot manipulators. When sampling in the search space, two ...constraints, namely, path length and manipulability measure, are imposed to find a minimal-cost path connecting the start and goal points. By tracking the generated path, a robot manipulator's end-effector can traverse the workspace with a shorter length and, meanwhile, avoid configuration singularities. A constrained closed-loop inverse kinematics technique is utilized to exploit the kinematic redundancy to assign a higher manipulability to an end-effector position. Additionally, the metrics of path length and manipulability measure are used to determine the adaptive step size for the RRT* planner. This helps the space-filling tree to grow efficiently toward unsearched areas and find an optimal path. Simulation analysis and experimental results of a six-degree-of-freedom FANUC-M-20iA industrial robot illustrate the efficiency of the proposed path planning methods.
In-depth studies of algorithms for solving motion planning problems have been conducted due to the rapid popularization and development of unmanned aerial vehicles in previous decades. Among them, ...the classic rapidly exploring random tree (RRT) algorithm has derivative algorithms (e.g., RRT*, Q-RRT*, and F-RRT*) that focus on the optimal path cost of the initial solution. Other improved algorithms, such as RRT-connect and BG-RRT, focus on the optimal time of the initial solution. This article proposes an improved density gradient-RRT (DG-RRT) algorithm based on RRT that improves the efficiency of the guide point and reduces the time lost in the process of obtaining the initial solution through the dynamic gradient sampling strategy. Simultaneously, it reduces the path cost by reconstructing the output path. The proposed algorithm is an expansion algorithm of a random tree, and the performance of the algorithm can be further improved by combining it with other RRT optimization algorithms. DG-RRT and other algorithms are compared in different environments through simulation experiments to verify the advantages of DG-RRT. In addition, it used a set of simulation flight tests to verify the feasibility of the DG-RRT algorithm for UAV path planning.
This article presents an algorithm termed as multiobjective dynamic rapidly exploring random (MOD-RRT*), which is suitable for robot navigation in unknown dynamic environment. The algorithm is ...composed of a path generation procedure and a path replanning one. First, a modified RRT* is utilized to obtain an initial path, as well as generate a state tree structure as prior knowledge. Then, a shortcuting method is given to optimize the initial path. On this basis, another method is designed to replan the path if the current path is infeasible. The suggested approach can choose the best node among several candidates within a short time, where both path length and path smoothness are considered. Comparing with other static planning algorithms, the MOD-RRT* can generate a higher quality initial path. Simulations on the dynamic environment are conducted to clarify the efficient performance of our algorithm in avoiding unknown obstacles. Furthermore, real applicative experiment further proves the effectiveness of our approach in practical applications.
This brief presents a trajectory planning algorithm for aerial vehicles traveling in 3-D space while avoiding obstacles. The nature of the obstacles can be, for example, radar detection areas, ...cooperating and non-cooperating vehicles, and so on. Thus, it is a complex trajectory planning problem. The proposed planner is based on the optimal rapidly exploring random tree (RRT*) algorithm. Artificial potential fields are combined with the RRT* algorithm to accelerate the convergence speed to a suboptimal solution by biasing the random state generation. The performance of this framework is demonstrated on a complex missile application in a heterogeneous environment. Indeed, since the air density decreases exponentially with altitude, the maneuverability of the aerial vehicle depending on aerodynamic forces also decreases exponentially with altitude. To face this problem, the shortest paths of Dubins-like vehicles traveling in a heterogeneous environment are used to build the metric. In the simulation results, this framework can find the first solution with fewer iterations than the RRT and the RRT* algorithm. Moreover, the final solution obtained within a given number of iterations is closer to an optimal solution regarding the considered criterion.
Multi-objective optimization problem (MOP) plays an increasingly important role in finance and engineering. In order to obtain more accurate and evenly distributed target solution set to a ...multi-objective programming, a novel swarm exploring neural dynamics (SEND) method is proposed, analyzed and applied in this paper. Specifically, a scalarization approach is firstly applied to transform the MOP into a group of subproblems. Secondly, each subproblem is solved by a varying parameter recurrent neural network (VP-RNN). By solving these problems, a group of Pareto optimal solutions are obtained. Thirdly, a population evolution weight optimization algorithm is used to diversify the solution set to obtain evenly distributed solutions. Simulation results demonstrate that the proposed SEND method can obtain a more accurate and evenly distributed solution set than some previous methods and the convergence rate is faster than the state-of-art methods, such as collaborative neurodynamic approach (CNA).
•A multi-start simultaneous exploration DRL method is proposed for process planning problems.•Designed a context vector for process planning problems.•A dynamic masking mechanism specific to the ...process planning problem is utilized.
Process planning is a vital stage in machining process, serving as the bridge between product design and manufacturing. However, due to the NP-hard nature of obtaining optimal or near-optimal process planning solutions under given manufacturing resources and constraints, process planning is a challenging problem. Moreover, existing methods struggle to address the issue of dynamic changes in machining resources during actual production. To overcome the challenges of finding optimal or near-optimal solutions in process planning problems and adapting to dynamic resource changes in actual production, this paper introduces a deep reinforcement learning (DRL) approach with multiple starting nodes exploration for process planning problems. Initially the objective function and constraints of process planning was transformed into matrix representations for subsequent DRL training. Subsequently, a specialized encoder-decoder architecture was designed for process planning, incorporating a dynamic mask mechanism to prevent generated solutions from violating constraints. Finally, a multi-start exploration policy gradient method was employed to train DRL, obtaining cost-effective and efficient process planning solutions. Experimental results demonstrated the effectiveness and superiority of the proposed method, contributing to the improvement of process planning efficiency in practical manufacturing.
