Leveraging Experience in Lifelong Multi-Agent Pathfinding

Many existing multi-agent path finding (MAPF) solvers focus on completeness, speed, or optimization. However, completeness and rapidity are usually in conflict with each other, which makes these algorithms far from satisfactory in practical applications. Motivated by this realistic requirement, we propose an efficient decoupling method to accelerate the solution of large MAPF problems. First, we define the concept of ‘non-essential vertex’-vertices which are not needed to solve a MAPF problem, and a scheme to identify them. Then, a decoupling scheme based on ‘non-essential vertex’ is proposed, which will assign higher priorities to agents whose goal positions are non-essential vertices and lower priorities to agents whose start positions are non-essential vertices. That is, invoking our decoupling algorithm can decouple any given MAPF problem into three subproblems while maintaining the completeness of the solution. All three sub-MAPF problems can be solved sequentially by a complete solver (e.g., CBS or EECBS, etc.), and two of them can also be solved by a prioritized planning algorithm. We have conducted several experiments in different workspaces, and the statistical results show that the proposed decoupling method significantly improves the speed and success rate of existing MAPF solvers with almost no degradation in solution quality when solving problems with high agent density. In addition, the solving efficiency can be further improved if the prioritized planning algorithm is invoked to solve the first and third sub-MAPF problems.

Multi-Agent Path Finding (MAPF) is an important problem with a variety of applications. Its aim is to find collision-free paths for agents having separate start and goal positions. This work proposes a new lexicographic bi-objective MAPF considering different task types, where agents are divided into two kinds to perform critical and acritical tasks. This is common in practical intelligent warehousing scenarios where a critical/acritical-task-performing agent (called c-agent and a-agent, respectively) may represent a full-load/no-load one or the one conducting urgent/non-urgent tasks. The primary objective is to minimize the sum-of-costs of c-agents, while the secondary objective is to minimize the sum-of-costs of a-agents. Two MAPF algorithms are modified to fit and solve the concerned problem for the first time. Moreover, four improvement strategies are embedded to the proposed algorithms and proved to be effective in solving MAPF problems with different task types.

In the multi-agent path finding problem (MAPF) we are given a set of agents each with respective start and goal positions. The task is to find paths for all agents while avoiding collisions, aiming to minimize a given objective function. Many MAPF solvers were introduced in the past decade for optimizing two specific objective functions: sum-of-costs and makespan. Two prominent categories of solvers can be distinguished: search-based solvers and compilation-based solvers. Search-based solvers were developed and tested for the sum-of-costs objective, while the most prominent compilation-based solvers that are built around Boolean satisfiability (SAT) were designed for the makespan objective. Very little is known on the performance and relevance of solvers from the compilation-based approach on the sum-of-costs objective. In this paper, we start to close the gap between these cost functions in the compilation-based approach. Our main contribution is a new SAT-based MAPF solver called MDD-SAT, that is directly aimed to optimally solve the MAPF problem under the sum-of-costs objective function. Using both a lower bound on the sum-of-costs and an upper bound on the makespan, MDD-SAT is able to generate a reasonable number of Boolean variables in our SAT encoding. We then further improve the encoding by borrowing ideas from ICTS, a search-based solver. In addition, we show that concepts applicable in search-based solvers like ICTS and ICBS are applicable in the SAT-based approach as well. Specifically, we integrate independence detection, a generic technique for decomposing an MAPF instance into independent subproblems, into our SAT-based approach, and we design a relaxation of our optimal SAT-based solver that results in a bounded suboptimal SAT-based solver. Experimental evaluation on several domains shows that there are many scenarios where our SAT-based methods outperform state-of-the-art sum-of-costs search-based solvers, such as variants of the ICTS and ICBS algorithms.

We address rule-based algorithms for multi-agent path finding (MAPF). MAPF is a task of finding non-conflicting paths connecting agents' initial and goal positions in a shared environment specified via an undirected graph. Rule-based algorithms use a fixed set of predefined primitive operations to move agents to their goal positions in a complete manner. We propose to apply spectral clustering on the underlying graph to decompose the graph into highly connected components and move agents to their goal cluster first before the rule-based algorithm is applied. The benefit of this approach is twofold: (1) the rule-based algorithms are often more efficient on highly connected clusters and (2) we can potentially run the algorithms in parallel on individual clusters.

A simple effective method for path planning based on a growing self-organizing elastic neural network, enhanced with a heuristic for the exploration of local directions is presented. The general problem is to find a collision-free path for moving objects among a set of obstacles. A path is represented by an interconnected set of processing units in the elastic self organizing network. The algorithm is initiated with a straight path defined by a small number of processing units between the start and goal positions. The two units at the extremes of the network are static and are located at the start and goal positions, the remaining units are adaptive. Using a local sampling strategy of the points around each processing unit, a Kohonen type learning and a simple processing units growing rule the initial straight path evolves into a collision free path. The proposed algorithm was experimentally tested for 2 DOF and 3 DOF robots on a workspace cluttered with random and non random distributed obstacles. It is shown that with very little computational effort a satisfactory free collision path is calculated.

Motor planning is a hierarchical process that is typically organized around an action's goal (e.g., drinking from a cup). However, the motor plan depends not only on the goal but also on the current body state. Here, we investigated how one's own body posture interacts with planning of goal-directed actions. Participants engaged in a grasp selection (GS) task while we manipulated their arm posture. They had to indicate how they would grasp a bar when transporting it from a start to goal position and orientation. We compared situations in which one's body posture was in-congruent with the start posture and/or goal posture of the planned movement. Behavioral results show that GS took longer when one's own body state was incongruent with the goal posture of the planned movement. Correspondingly, neural activity in the intraparietal sulcus (IPS) and extrastriate body area (EBA) was modulated by congruency between the body state and the action plan. IPS was sensitive to overall congruency between body posture and action plan, while the EBA was sensitive specifically to goal posture congruency. Together, our results suggest that IPS maintains an internal state of one's own body posture, while EBA contains a representation of the goal posture of the action plan.

최근 HFC(Hybrid Fiber Coax) CATV 망이 초고속망의 하부구조로 자리를 잡아가고 있다. HFC-CATV 망의 상향채널은 가입자에서 헤드엔드로 데이터 전송시 500가입자 이상이 공유하기 때문에 충돌이 발생하게 된다. 특히, 음성, 비디오 그리고 데이터와 같은 다양한 종류의 트래픽 QoS(Qualify of Service)를 만족시키기 위해서 서비스별로 우선순위를 고려한 알고리즘이 요구되어진다. IEEE 802.14에서 제안된 우선순위 기반 3진 트리(Ternary tree)알고리즘은 정적 PNA(Priority New Access) 슬롯 구조를 가진다. 이 방법은 실시간으로 변화하는 트래픽의 요구 지연(request delay)이 높게 유지되며 서로 다른 우선순위 트래픽의 요구 경쟁 확률이 동일하다는 문제점을 가지고 있다. 따라서, 본 논문에서는 동적 PNA(Priority New Access) 슬롯 구조를 제안하여 높은 우선순위 트래픽의 요구지연을 낮게 하고 새로운 스테이션에 대해 우선순위 별로 경쟁 확률을 다르게 하는 알고리즘을 제안한다. 동일한 조건하에 두 알고리즘에 대해 요구지연 값을 비교 분석하였다. 성능 분석 결과 동적 PNA 우선순위 충돌해결알고리즘이 정적 PNA 우선순위 충돌해결알고리즘보다 낮은 요구지연을 가진다는 것을 확인할 수 있었다. Recently, the HFC-CATV network stand in a substructure of superhighway information network. Because of sharing up to 500 of subscribes, the Collision Resolution Algorithm needs in the upstream channel of HFC-CATV network. In order to provide Quality of Service (QoS) to users with real-time data such as voice, video and interactive service, the research of Collision Resolution Algorithm must include an effective priority scheme. In IEEE 802.14, the Collision Resolution Algorithm has high request delay because of static PNA(Priority New Access) slots structure and different priority traffics with the same probability. In order to resolve this problem, this paper proposed dynamic priority collision resolution algorithm with ternary tree algorithm. It has low request delay according to an increase of traffic load because high priority traffic first resolve and new traffic content with different probability. In the result of the simulation, it demonstrated that the proposed algorithm needs lower request delay than that of ternary tree algorithm with static PNA slots structure.

The task in the multi-agent path finding problem (MAPF) is to find paths for multiple agents, each with a different start and goal position, such that agents do not collide. A successful optimal MAPF solver is the conflict-based search (CBS) algorithm. CBS is a two level algorithm where special conditions ensure it returns the optimal solution. Solving MAPF optimally is proven to be NP-hard, hence CBS and all other optimal solvers do not scale up. We propose several ways to relax the optimality conditions of CBS trading solution quality for runtime as well as bounded-suboptimal variants, where the returned solution is guaranteed to be within a constant factor from optimal solution cost. Experimental results show the benefits of our new approach; a massive reduction in running time is presented while sacrificing a minor loss in solution quality. Our new algorithms outperform other existing algorithms in most of the cases.

The IP address lookup in the communication link technologies is becoming a major bottleneck by exploring the exact matching and the parallel hashing in each prefix length. In this paper, we present an algorithm for IP address lookup and collision resolution to reduce memory space and solve collision for high speed routing based on multiple table indexing structure. Our scheme makes possible direct access in a designated IP prefix without replicating the forwarding table entries, and provides efficient collision resolution for fast route lookup with only small memory. Simulation results indicate that our algorithm provides a faster lookup than the conventional lookup schemes.

The performance of hybrid random-access systems (HRAS) which use a combination of multiuser codes (MUC) and collision resolution algorithms (CRA) to accommodate the bursty transmissions of many independent users on a single communication channel is analyzed. Besides the computation of the system throughput, another contribution of this paper is the determination of the properties which a MUC must possess such that resulting HRAS performs better than a system which is only based on CRA'S.

The subject of this paper is the computation of paths for mobile robots that navigate from a start position to a goal position in environments with static obstacles. Specifically, we focus on paths that are represented by straight lines. Such paths can for example directly be followed by omni-directional robots or can be used as an initial solution for path smoothing. In this context, the most common performance metrics are the path length, the obstacle clearance and the computation time. In this paper, we develop a new path planning algorithm that addresses all the stated performance metrics. Our method first determines all possible connections between the start position and goal position along the edges of the generalized Voronoi diagram (GVD) of a given obstacle map. The shortest connections are then refined using a balanced method for creating shortcuts along existing waypoints and introducing new waypoints in order to cut corners. As an important feature, our method reduces the number of required waypoints by iteratively adding new waypoints and then removing unnecessary waypoints along solution paths. Moreover, our method takes into account multiple start-goal connections, since the shortest start-goal connection along the edges of the GVD might not lead to the shortest solution path. A comprehensive computational evaluation for a large number of maps with different properties shows that the proposed method outperforms sampling-based algorithms such as Probabilistic Roadmaps (PRM) and exact methods such as Visibility Graphs (VG) by computing close-to-optimal solution paths with a specified minimum obstacle clearance in less time.

Robust fast continuous path tracking control is the important technology for the multi-degrees-of-freedom position control system such as X-Y table, robot manipulator and so on. Especially, large acceleration and/or deceleration torque is necessary for fast continuous path tracking control at its start position and its goal position. Each motor of high speed path tracking control has the torque limitation. This motor cannot always obtain the enough acceleration and/or deceleration torque due to its torque limitation. Hence, in controlling a multi-degree-of-freedom position control system, even if motor torque of only one axis has torque limitation, the actual position response is not equal to the trajectory reference. In fast continuous path tracking control of multi-degree-of-freedom position control system, the position error sometimes increases according to the unknown disturbance torque and inertia variation. In order to overcome these problems, this paper proposes a new robust fast continuous path tracking control algorithm taking into account both the torque saturation and the coordinated motion, which is based on the robust acceleration control using disturbance observer. The proposed method assures the coordinated motion considering the torque limitation. The effectiveness of the proposed method is confirmed by the experimental results in this paper.

Any obstacle-free path planning algorithm, in general, gives a sequence of waypoints that connect start and goal positions by a sequence of straight lines, which does not ensure the smoothness and the dynamic feasibility to maneuver the MAV. Kinodynamic-based motion planning is one of the ways to impose dynamic feasibility in planning. However, kinodynamic motion planning is not an optimal solution due to high computational demands for real-time applications. Thus, we explore path planning followed by kinodynamic smoothing while ensuring the dynamic feasibility of MAV. The main difference in the proposed technique is not to use kinodynamic planning when finding a feasible path, but rather to apply kinodynamic smoothing along the obtained feasible path. We have chosen a geometric-based path planning algorithm “RRT*” as the path finding algorithm. In the proposed technique, we modified the original RRT* introducing an adaptive search space and a steering function that helps to increase the consistency of the planner. Moreover, we propose a multiple RRT* that generates a set of desired paths. The optimal path from the generated paths is selected based on a cost function. Afterwards, we apply kinodynamic smoothing that will result in a dynamically feasible as well as obstacle-free path. Thereafter, a b-spline-based trajectory is generated to maneuver the vehicle autonomously in unknown environments. Finally, we have tested the proposed technique in various simulated environments. According to the experiment results, we were able to speed up the path planning task by 1.3 times when using the proposed multiple RRT* over the original RRT*.

This paper considers the collision resolution algorithm when multiple packets transmit simultaneously in each slot in a non-orthogonal multiple access (NOMA) systems. Packets are divided into two groups and two different received powers can be selected. Packets in each group can select one power level as their received power. As more than one packet may transmit in a time slot in each group, collision will happen and throughput may decrease at the same time. We introduce the collision resolution period (CRP) algorithm and the collided packets will transmit in the CRP. By controlling the transmission probability in each group, we can make a big different comparing to the system without a CRP. This paper will show the optimal transmission probability and the maximum throughput. Numerical results show that our proposed algorithm yields the maximum throughput of 0.6489.

We consider connectionless packet switched CDMA systems where all users have to share a common bandwidth and signature sequence on a contention basis. Multiple access is achieved by the simultaneous acquisition of different users' spread spectrum transmissions. The throughput of a pure random access system of this kind suffers from acquisition errors and timing mismatches of the active users as well as instability issues. We consider schemes that use information available at the physical layer, i.e., at the output of the multiuser access detector (MUAD) designed for this random access system, to improve and stabilize the throughput. We consider timing randomization, pseudo-Bayesian stabilization and collision resolution algorithms and report the corresponding numerical results.