Yasir Javed

COROS: A Multi-Agent Software Architecture for Cooperative and Autonomous Service Robots

Building distributed applications for cooperative service robots systems is a very challenging task from software engineering perspective. Indeed, apart from the complexity of designing software components for the control of a single autonomous robot, cooperative multi-robot systems require additional care in the design of software components to ensure communication and coordination between the robotic agents. This chapter proposes COROS, a new multi-agent software architecture for cooperative and autonomous service robots with the objective to make easier the design and development of multi-robot applications. We present a high-level conceptual architecture for multi-agent robotics systems that represents a generic framework for cooperative multi-robot applications. Furthermore, we present an instantiation of this generic architecture with an implementation software architecture on top of the Robot Operating System (ROS) middleware. The proposed concrete software architecture follows a component-based approach to ensure modularity, software reuse, extensibility and scalability of the multi-robot operational software. In addition, one major added value of our architecture is that it provides a tangible solution to supporting multi-robot software development for the ROS middleware, as ROS was originally designed for single-robot applications. We also demonstrate a sample of real-world case studies of cooperative and autonomous service robots applications in an office-like environment, including discovery and courier delivery applications.

Turtlebot at Office: A Service-Oriented Software Architecture for Personal Assistant Robots Using ROS

This paper presents the design of an assistive mobile robot to support people in their everyday activities in office and home environments. The contribution of this paper consists in the design of a modular component-based software architecture that provides different abstraction layers on top of Robot Operating System (ROS) to make easier the design and development of service robots with ROS. The first abstraction layer is the COROS framework composed of complementary software subsystems providing different interfaces between ROS and the client applications. The second abstraction layer is the integration of Web services into ROS to allow client applications to seamlessly and transparently interact with the robot while hiding all implementation details. The proposed software architecture was validated through a experimental prototype of Turtlebot deployed in University campus. Furthermore, we outline the challenges incurred during experimentation and focus on lessons learned throughout the implementation and deployment.

Open source web application security: A static analysis approach

In this paper, we have tested several open source web applications against common security vulnerabilities. These vulnerabilities spans from unnecessary data member declaration to leaving gaps for SQL injection. The static security vulnerabilities testing was done in three categories (1) Dodgy code vulnerabilities (2) Malicious code vulnerabilities (3) Security code vulnerabilities on seven (7) different web applications built in Java. It is evident from the obtained results that almost all selected applications have similar kind of vulnerabilities that might have been introduced due to hasty programming or lack of developer knowledge against security vulnerabilities. We recommend to create an intelligent development framework that can provide suggestions for secure development by overcoming common vulnerabilities, can add missing code and can learn from expert developers practices to overcome the security vulnerabilities.

Robot Path Planning and Cooperation - Foundations, Algorithms and Experimentations

This book presents extensive research on two main problems in robotics: the path planning problem and the multirobot task allocation problem. It is the first book to provide a comprehensive solution for using these techniques in large-scale environments containing randomly scattered obstacles. The research conducted resulted in tangible results both in theory and in practice. For path planning, new algorithms for large-scale problems are devised and implemented and integrated into the Robot Operating System (ROS). The book also discusses the parallelism advantage of cloud computing techniques to solve the path planning problem, and, for multi-robot task allocation, it addresses the task assignment problem and the multiple traveling salesman problem for mobile robots applications. In addition, four new algorithms have been devised to investigate the cooperation issues with extensive simulations and comparative performance evaluation. The algorithms are implemented and simulated in MATLAB and Webots. Studies in Computational Intelligence

Different Approaches to Solve the MRTA Problem

The multi-robot task allocation problem is a fundamental problem in robotics research area. The problem roughly consists of finding an optimal allocation of tasks among several robots to reduce the mission cost to a minimum. As mentioned in Chap. 6, extensive research has been conducted in the area for answering the following question: Which robot should execute which task? In this chapter, we design different solutions to solve the MRTA problem. We propose four different approaches: an improved distributed market-based approach (IDMB), a clustering market-based approach (CM-MTSP), a fuzzy logic-based approach (FL-MTSP), and Move-and-Improve approach. These approaches must define how tasks are assigned to the robots. The IDBM, CM-MTSP, and Move-and-Improve approaches are based on the use of an auction process where bids are used to evaluate the assignment. The FL-MTSP is based on the use of the fuzzy logic algebra to combine objectives to be optimized.

Performance Analysis of the MRTA Approaches for Autonomous Mobile Robot

The multi-robot task allocation is a fundamental problem in robotics research area. Indeed, robots are typically intended to collaborate together to achieve a given goal. This chapter studies the performance of the IDBM, CM-MTSP, FL-MTSP, and Move-and-Improve approaches. In order to highlight the performance of the proposed schemes, we compared each one to appropriate existing ones. IDMB was compared with the RTMA [1], CM-MTSP was compared with single-objective and greedy algorithms, and FL-MTSP was compared with a centralized approach based on genetic algorithm and with NSGA-II algorithm. To validate the efficiency of the Move-and-Improve distributed algorithm, we first conducted extensive simulations and evaluated its performance in terms of the total traveled distance and the ratio of overlaped targets under different settings. The simulation results show that IDMB and Move-and-Improve algorithms produce near-optimal solutions. Also, CM-MTSP and FL-MTSP provide a good trade-off between conflicting objectives.

Introduction to Mobile Robot Path Planning

Robotic is now gaining a lot of space in our daily life and in several areas in modern industry automation and cyber-physical applications. This requires embedding intelligence into these robots for ensuring (near)-optimal solutions to task execution. Thus, a lot of research problems that pertain to robotic applications have arisen such as planning (path, motion, and mission), task allocation problems, navigation, tracking. In this chapter, we focused on the path planning research problem.

General Background on Multi-robot Task Allocation

Multi-robot systems (MRSss) face several challenges, but the most typical problem is the multi-robot tasks allocation (MRTA). It consists in finding the efficient allocation mechanism in order to assign different tasks to the set of available robots. Toward this objective, robots will work as cooperative agents. MRTA aims at ensuring an efficient execution of tasks under consideration and thus minimizing the overall system cost. Various research works have solved the MRTA problem using the multiple traveling salesman problem (MTSP) formulation. In this context, an overview on MRTA and MTSP is given in this chapter. Furthermore, a summary of the related works is presented.

Background on Artificial Intelligence Algorithms for Global Path Planning

In the literature, numerous path planning algorithms have been proposed. Although the objective of these algorithms is to find the shortest path between two positions A and B in a particular environment, there are several algorithms based on a diversity of approaches to find a solution to this problem. The complexity of algorithms depends on the underlying techniques and on other external parameters, including the accuracy of the map and the number of obstacles. It is impossible to enumerate all these approaches in this chapter, but we will shed the light on the most used approaches in the literature.

Design and Evaluation of Intelligent Global Path Planning Algorithms

Global path planning is a crucial component for robot navigation in map-based environments. It consists in finding the shortest path between start and goal locations. The analysis of existing literature in Chap. 2 shows two main approaches commonly used to address the path planning problem: (1) exact methods and (2) heuristic methods. A* and Dijkstra are known to be the most widely used exact methods for mobile robot global path planning. On the other hand, several heuristic methods based on ant colony optimization (ACO), genetic algorithms (GA), Tabu Search (TS), and hybrid approaches of both have been proposed in the literature. One might wonder which of these methods is more effective for the robot path planning problem. Several questions also arise: Do exact methods consistently outperform heuristic methods? If so, why? Is it possible to devise more powerful hybrid approaches using the advantages of exact and heuristics methods? To the best of our knowledge, there is no comprehensive comparison between exact and heuristic methods in solving the path planning problem. This chapter fills the gap, addresses the aforementioned research questions, and proposes a comprehensive simulation study of exact and heuristic global path planners to identify the more appropriate technique for the global path planning.