Path Planning for Agent Sensors in Indoor Environment Using Bio-inspired Routing with Dijkstra Algorithm

Abstract

Wireless Sensor Networks (WSN) has been a trending area of research for a while, thanks in part to rapid growth of Internet of Things (IoT). A WSN includes a number of independent, self-sustainable agents that collaborate with each other to sense information from the environment and perform a specific task. However, due to the agents' limited resources, optimization techniques are essential for the efficient performance in these systems. In this thesis, a simulation study was designed and implemented that presents indoor unknown environments to solve search and rescue problems in s semi-distributed fashion. This relied on implementing mazes which could be close to real life situations where walls represent obstacles that agents should avoid. Agents are injected in the maze and they start moving following a bio-inspired technique, such as the Inverted Ant Algorithm (iAA) to find a target which is a device emitting signal in the uncertain environment and, crucially for this Thesis, whenever it is located by any agent, a path should be created back to the source. This path is optimized using Dijkstra Algorithm by constructing a graph with all agents and weighted edges avoiding crossing walls and considering transmission range. The situation with penetrable walls is considered, because signals can be transmitted through different materials. Moreover, a novel proposed algorithm was implemented to solve the issue of the path loss. The performance of the aggregated system solution was studied and the study of the effect of changing parameters on the system was done as well. After explaining how the target is found using bio-inspired algorithm (a version of Ant Algorithm), the main research questions that this thesis answers are: (a)How to construct the graph of agents' locations after finding target with weighted edges considering transmission range and obstacles? (b) How to find the optimal shortest path of agents relies back to sources and implements it in the simulation? (c) How to avoid the issue of path loss?

Date of Award	Dec 2018
Original language	American English