Abstract

Autonomous robots can be used in a decentralized environment to contain threats. While working in this capacity, these motor propelled robots are constantly moving, therefore drawing a large amount of current from the battery. If these algorithms are to be implemented in hardware, it is important to ensure that the robots move only when necessary in an effort to optimize battery life. This work introduces static wireless sensors to assist robots in detecting threats. By having a sufficient number of wireless sensors available to detect threats, it is hypothesized that a similar containment performance can be achieved with less robot movements. When not actively containing threats, the robots may enter a sleep mode thus optimizing energy conservation. The notion of multimode operations has been utilized in other wireless sensor network applications. In the field of cooperative robotics, however, little has been investigated for system performance when both mobile robots and static sensors coexist. This work leverages previously developed multi-threat containment algorithms and the notion of multi-mode operations from wireless sensor network research community and examines the scenarios where wireless sensors can benefit the overall system performance. Battery models and additional sensor and obstacle objects are introduced to a previously developed simulator, MAHESHDAS. Various battery models and parameters are considered to mimic a realistic environment. The sensor nodes occupy a small amount of physical space and therefore assist the robots while also limiting their movements. Robots are assumed to be ground vehicles, and will need to avoid collisions of each other as well as sensor nodes and other obstacles. Repulsion forces are used to model the collision avoidance between the various objects. The percentage of threats contained, the time to contain threats, and the average robot lifetime are compared in different operational scenarios. The simulation results demonstrate that the introduction of wireless sensors improve the average robot lifetime when the threats do not occur too often and when the sensor repulsion force is relatively small. Uniform sensor placement is also shown to perform better than random deployment.

Library of Congress Subject Headings

Wireless sensor networks; Autonomous robots--Control systems--Computer simulation; Multisensor data fusion

Publication Date

8-1-2009

Document Type

Thesis

Department, Program, or Center

Computer Engineering (KGCOE)

Advisor

Yang, Shanchieh

Comments

Note: imported from RIT’s Digital Media Library running on DSpace to RIT Scholar Works. Physical copy available through RIT's The Wallace Library at: TK7872.D48 E55 2009

Campus

RIT – Main Campus

Download

Share

COinS