A GPS-Less Localization and Mobility Modelling (LMM) System for Wildlife Tracking

Existing wildlife tracking solutions typically use sensor nodes with specialised facilities, such as long-range radio, solar array of cells and Global Positioning System (GPS). This introduces additional manufacturing cost, increased energy and memory consumptions and increased sensor node weight. This paper proposes a novel Localization and Mobility Modelling (LMM) system, that can carry out wildlife tracking by merely using low-cost, lightweight sensor nodes and using short-range peer-to-peer communication facilities only, i.e. without the need for any specialised facilities. This is done by using two computationally simple operations, which are: (i) aggregated data collections from sensor nodes via peer-to-peer communications in a distributed manner, and (ii) estimation of sensor nodes & x2019; movement traces using trilateration. The computational load placed on each sensor node is just that of data collection and aggregation, whereas movement traces estimation is carried out on a backend server, separated from the sensor nodes. In the design of the LMM system, we have: (i) carried out an empirical evaluation of different parameter value settings for data collection to develop a Multi-Zone Multi-Hierarchy (MZMH) communication structure, (ii) demonstrated a novel use of an Aggregation based Topology Learning (ATL) protocol for collecting sensor nodes & x2019; topology data using peer-to-peer multi-hop communications, and (iii) used a novel Location Estimation (LE) method for estimating sensor nodes & x2019; movement traces from the collected topology data. The evaluation results show that the LMM system can accurately estimate sensor nodes & x2019; movement traces but with significantly less energy and memory costs, demonstrating its cost-efficiency as compared to the related wildlife tracking solutions.