WILDSENSING: A Hybrid Framework of Mobile and Sensor Nodes for Wildlife Monitoring

Lead Research Organisation: University of Oxford
Department Name: Computer Science

Abstract

Technological advances in Micro-Electro-Mechanical Systems (MEMS) are envisaged to allow the dense deployment of nodes with sensing, communication and processing capabilities in large areas for monitoring purposes. In this project we offer an alternative to plain multi hop data forwarding through the sensor network. Our approach suggests the forwarding of sensor data and its storage in selected nodes (storage nodes) from where the data will be collected later on by roaming mobile nodes. This new operational setting will leverage recent advances in mobile technology to relieve the sensor network from heavy multi-hop communication tasks. It will exploit the vast availability of a variety of different mobile devices (e.g., phones, pdas and domain specific wireless-equipped devices such as health monitors) and the potential for user or unmanned vehicle mobility. Mobile devices are equipped with one or more wireless network interfaces (Bluetooth, 802.11 etc), which makes them able to connect and interact with storage nodes in radio range, in an ad hoc manner. An application that would particularly benefit from continous monitoring using sensor nodes is wildlife monitoring. Zoologists would be able to detect social behavior patterns of wild animals (e.g. animal movement patterns), in combination with microclimate conditions, to protect the animals' habitat and ensure their well-being. Current approaches to wildlife monitoring and conservation often still rely on labour intensive techniques for making observations of animal behaviour or for tracking animal movements with established (but outmoded) VHF telemetry equipment. The typical mode of monitoring is to send staff to every single sensor node in the field, to collect sensor readings. The raw data is collected by staff, brought together in a lab, and processed in a centralized manner. The heavy reliance on field-staff for animal monitoring currently incurs considerable employment costs and overheads for ancillary equipment. The use of personnel working alone at night in forests also has significant health and safety implications, and the scrutiny of the Health and Safety Executive is likely to jeopardise many of these protocols in the future.
 
Description WildSensing is a cross disciplinary project focusing on monitoring wild badgers in their natural habitat in Wytham Woods, using a mixture of static and mobile sensing technology.

Current approaches using VHF beacons are highly labour intensive and involve night-time tracking. GPS tracking collars typically perform poorly under heavy tree cover. The WildSensing approach was to leverage the power of sensor network technology to provide continuous and long term monitoring. Over 40 animals were equipped with tracking collars which periodically emit radio beacons. In order to rapidly deploy the devices, existing active RFID devices from WaveTrend were used. These transmissions were detected by a network of 30 sensor nodes placed at strategic locations, such as at setts. In a one year operational period, over 26 million transmissions were detected and logged. The deployment experience has led to the development of algorithms to deliver data using a hierarchical delay tolerant approach, surprising results about the effect of rainfall on link behaviour, and hardware and firmware advances. A biologically inspired method of evolving code for distributed networks using an abstraction of Gene Regulation was also devised. The initial data has revealed some interesting badger behaviour which has previously been speculated upon, but never before been observed directly, such as dispersal of animals from one sett to another.

Another avenue of research that has been recently explored is localizing badgers when they are underground. Radio signals are unable to penetrate soil, and thus conventional tracking modalities cannot be used. Instead, a novel method using low frequency magnetic fields was developed. As a byproduct of localizing animals over time, the tunnel structure itself can be revealed. This new research area is likely to lead to some interesting discoveries about badger social structure underground, something which has been impossible to do until now.
Exploitation Route They can be of great value to zoologists interested in understanding how animals behave and interact with each other, underground and above ground, and in different environmental conditions.
Sectors Environment
URL https://www.cs.ox.ac.uk/projects/WILDSENSING/