Tang Zhenmin

Control Architecture for Autonomous Multi-Robot System: Survey and Analysis

The basic consideration of autonomous multi-robot system (MRS) control architecture is dealing with the whole system as a group, in which the individual robots interact with each other, avoid conflict, and cooperate to accomplish the tasks reliably and efficiently in complex environment. This paper systematically surveys and summarizes related studies on control architectures for multi-robot systems from definition and evolutions to theories and technologies corresponded to research field. Some representational demonstrations and instances are analyzed. Finally the future directions of autonomous MRS architecture are overlooked in the end.

Complex task query based on cloud computing Resource: Improved approximate skyline algorithm

Cloud computing aims at sharing all resource completely, which include CPUs computing ability, store capacity, network bandwidth and etc. Most cloud computing users task requisitions are among multiple target requisition which are hard to be met by these requisition conditions simultaneously. So it is the key to find out the most approximate node machine among large number of node machines in cloud computing model. This paper presents an improved approximate skyline algorithm to hunt to most approximate node machine economically, and provides an alternative reference for the coming task allocation.

Application of RBF neural network based on AP clustering in engine fault diagnosis

RBF neural network is widely used in intelligent fault diagnosis with its good performance for nonlinear problems. But the nodes number in hidden layer is difficult to get, so the advanced RBF neural network (AP-RBF) based on AP clustering is proposed to gain proper hidden layer efficiently. In AP-RBF, the exemplars obtained by AP clustering are used to construct hidden layer of RBF network. The results of engine fault diagnosis show that AP-RBF can achieve higher accuracy through more compact hidden layer than traditional RBF and RBF based on subtractive clustering (C-RBF).

Uncertainty measurement for environmental recognition ability of intelligent robot

Ground intelligent robot may fall across some unpredictable tasks and environmental conditions. Traditional methods on ability evaluation barely concern fuzziness and randomness of system. This paper presents a method with uncertainty measurement theory to evaluate the environmental recognition ability of intelligent robot. Firstly, a cloud model with parameters including expectation, entropy and hyper-entropy is proposed. Then, a Two-Dimensional Multi-rules Cloud Reasoning algorithm (2DMCR) is addressed basing the cloud model. With a set of X-condition reasoning rules on Environmental Recognition ability, the algorithm constructs a multi-rules generator and calculates the expectation of cloud model. Finally, a practical experiment based on ground intelligent robots in our laboratory is given. The results show that this method is feasible in determining appropriate level of ER ability and can also be used in evaluation of other intelligent system as well.

Particle swarm based stereo algorithm and disparity map evaluation

In this paper, a new particle swarm based stereo algorithm is presented. Our motivation is to improve the accuracy of the disparity map by removing the mismatches caused by both occlusions and false targets. In our approach, the stereo matching problem is divided into two steps, including partial matching of segmented image and particle swarm optimization of the rest. The algorithm first takes advantage of SAD and Dynamic Programming to remove the mismatches mainly caused by visibility problems; after the first step, the algorithm selects all the rest image segmented regions, takes them as a particle and uses particle swarm to optimization it. In the second step, the cost function is defined on the pixel level, as well as on the segmented level, while the pixel level measures the data similarity based the current disparity map, the segmented level incorporates a smooth term. Results obtained for benchmark indicate that the proposed method is able to get rather accurate disparity maps.

Building configuration space for multiple UGVs

Classical configuration space (C-space) is often used in such a assumption that obstacles are static and robots have global knowledge of environment. Recently, there has been much more attention in multi-robot systems (MRS), unfortunately a MRS which have local sensing cannot meet the assumption. If it can be solved, the motion planning of the MRS benefits from previous motion planning algorithms in C-space. Toward resolving aforementioned problems, in the first place, we model the environment and robots as some sets of polygons, robots are unmanned ground vehicles (UGVs) with limited sensing range and can communicate with each other. Then a decentralized method is introduced by which two autonomous robots search environment and build C-space obstacles, while exchange and merge partial C-space generated by another robot every certain time interval until the whole C-space obstacles are established, it is similar to concurrent mapping and localization (CML), but we mainly focus on the algorithm of generating and merging C-space. Finally a simulated implementation demonstrates validity of our algorithm.