Multiple UAV search using agent based negotiation scheme

Abstract

Abstract-In this paper, we present an agent based negotiation scheme for multiple UAVs performing search operation on an unknown region. The UAVs are subjected to limited sensor range and can communicate with their neighbouring UAVs only. The UAVs use negotiation as the decision making mechanism for obtaining search paths. The scheme is scalable to large number of UAVs without much computational burden. We study the performance of uncertainty reduction strategies using negotiation schemes with different levels of information structures. The simulation results show that the search based on negotiation scheme with various kinds of information structures outperforms random and greedy strategies with identical information structures. I. INTRODUCTIONThe use of unmanned aerial vehicles for search and surveillance operations in unknown and hostile regions is becoming a reality. Coordinating these aerial vehicles, which perform the operation autonomously, is a difficult task. Usually, these vehicles have limited communication and decision-making capability. With large number of agents the computational overhead on the decision-making and coordination mechanism becomes high. In this paper, we present an agent-based negotiation scheme that scales well with increase in number of agents/vehicles and demands modest computational time.Cooperative search using multiple vehicles has wide variety of applications, such as search, surveillance, disaster management, task allocation, sensor networks, etc, and has attracted the attention of several researchers in recent times. In [1], [2], the authors address the problem of searching an unknown region with multiple vehicles cooperatively. A recursive approach is used for cooperative search using a multi-objective cost function, and q-step path planning. In [3], a team of UAVs is given the task of searching a region with unknown opportunities and hazards with an objective that the team has to maximize the regions of opportunity visited, while minimizing visits to regions of hazard, subject to constraints that the UAVs must remain connected by a communication network at all times and avoid collisions among themselves. Coordinating large number of UAVs for wide area search munitions using an agent architecture is discussed in [4], where the authors use the concept of teamwork in multi-agent systems for team plan sharing and instantiation, role allocation for the UAVs, and realtime onboard information updates. In [5], a knowledgebased framework is proposed for multiple agents performing

Communication-aware distributed PSO for dynamic robotic search

“…Unattended search drones are a particularly useful application for robots capable of these tasks [33][34][35]. These drones are intended for use over a wide area and are often assigned to the task of finding a mobile target.…”

Multiple vehicle Bayesian-based domain search with intermittent information sharing

“…Now assume that ǫ is a small number in the neighborhood of zero, under Sensing Condition 1, Π j∈Ni(t) ( 1 β j − 1)ǫ is also a small number close to zero. Hence, Equation (6) can be approximated as…”

“…The objective is to minimize the environment uncertainty in a finite amount of search time. With the same goal, in [6], the authors present an agent-based negotiation scheme for a multi-UAV search operation with limited sensory and communication ranges. In [7], the authors use the Modified Bayes Factor to model the level of confidence of target existence for an UAV search task in an uncertain environment.…”

Information merging in multi-UAV cooperative search

“…UAVs exchange messages to maintain proper distance among them in order to cover the whole search region uniformly. Negotiation among UAVs [23] is another approach that has been used to reduce the uncertainty about the target existence and to avoid covering the same area by multiple UAVs simultaneously. In a similar method, neighboring UAVs exchange proposals and mutually decide using self-assessment based decision making to cover a specific sub-region [24].…”