Research On Distributed Cooperative Location Based On UWB Ranging

With the advent of the 21st century,the sensor technology developing continuously,and the number of sensor nodes also gradually expand.Every two sensor nodes can communicate with each other.All the nodes constitute the Wireless Sensor Network(WSN),it is usually deployed in certain circumstances,used to perform different functions,such as traffic management,environmental monitoring,etc.In wireless sensor networks,the location information is vital.However,due to the complexity of the environment,the sensor nodes often lose GPS signals and the cost of GPS device is high,and the interaction of nodes is not used.So how to make full use of the synergistic effect of each node,co-location,is currently a hot research topic.Current existing positioning algorithms can be divided into:based on whether the positioning algorithm needs distance information between nodes,it can be divided into range-free algorithm and range-based algorithm.The former is simple and easy to implement but the positioning is not accurate,while the latter uses distance information and has higher accuracy.Similarly,according to whether the positioning algorithm needs to carry out complex operations in a central node,it can be divided into centralized algorithm and distributed algorithm.The centralized positioning algorithm requires the central node to complete the positioning calculation,which has higher accuracy but high time consumption and complexity.In the distributed localization algorithm,the computing task is distributed to each node,which has low complexity and less energy consumption.Therefore,how to design high precision and low complexity distributed co-location algorithm is an important challenge in wireless sensor network localization.Currently,there are three important problems in the mainstream distribution-based collaborative location scheme:(1)The clock frequency of ranging device is not accurate,lead to larger range error(2)In a certain distance between the two sensors to measure,we often use the middle range of multiple sensors multiple hops from additive to replace the Euclidean distance,lead to poor positioning after calculating precision(3)In the iterative method,we generally need to use the anchor nodes.Without anchor nodes,we can not locate nodes in iterating.In order to solve these problems,this paper designs a distributed co-location algorithm based on UWB ranging to improve accuracy and reduce complexity.The main work is summarized as follows:1.This paper studies the signal characteristics of UWB,which has the advantages of anti-multipath interference and high ranging accuracy,and studies the ranging mechanism.An improved time-of-flight asymmetric two-way ranging(BTOF-TWR)algorithm is proposed,and an error model based on time-of-flight is introduced,but the error caused by the processing time of the equipment is eliminated,and the error of overall accuracy is improved.The experimental system was built,and the system error was eliminated by polynomial fitting.Finally,when the real distance was 3m and 6m,the accuracy of BTOF-TWR algorithm was 2.98%and 5.92%higher than that of SDS-TWR algorithm,respectively.2.This paper proposes an improved MDS(MDS-IRBE)algorithm based on centroid coordinates.This algorithm is divided into the following three research sections:First,the topology of reference node is reconstructed by using the dissimilarity matrix to obtain its relative coordinates,so as to solve the problem that the anchorless nodes cannot be located;Secondly,according to the centroid model,the matrix equation of the relation between other nodes and reference nodes is given,which solves the problem that the multi-hop distance replaces the Euclidean distance.The third is the introduction of Richardson iteration,which makes the algorithm distributed,reduces the complexity of positioning.And the distributed positioning algorithm flow is proposed.In this paper,through a series of simulation,this paper first proves the convergence of the algorithm,and find the fastest convergence factor,and then design different scenario simulation.For randomly distributed nodes and uniformly distributed nodes in two-dimensional space.Using MDS-IRBE algorithm proposed in this paper,compared with local-MDS algorithm,the positioning accuracy improves by 8.36%and 4.78%respectively.Finally,this paper builds a collaborative positioning system and conducts positioning experiments on actual scenes.It is concluded that the accuracy of the proposed distributed positioning algorithm(MDS-IRBE)is 50.25%higher than that of the MDS algorithm.Since the MDS algorithm does not involve iteration,the positioning accuracy of the proposed algorithm has been significantly improved,which also strongly proves its effectiveness and feasibility.