Digital Battlespace

DSTL's Stone Soup to aid sensor fusion

18th June 2019 - 10:00 GMT | by Kate Martyr in London


Save this for later

The UK’s Defence Science and Technology Laboratory (DSTL) is hoping its novel approach to algorithm solving will improve tracking and fusion systems as part of its continued research into future situational awareness.

Stone Soup, a software framework that allows code sharing and algorithm comparison for the core mathematical components of situational awareness systems, was released as an open-source collaborative project in May 2019 under the DSTL Future Sensing and Situational Awareness (FSSA) programme.

The framework means that members of the Five Eyes nations are able to upload and test new components or compare tracking algorithms alongside accepted or state-of-the-art algorithms.

Industry is also able to insert its own data into the framework and run this against the standard suite of tracking algorithms it already contains.

It is hoped a collaborative approach will lead to better mathematical algorithms for sensor fusion which ‘is a mathematically challenging task that can be difficult for even the most powerful computers to perform,’ a DSTL spokesperson told Shephard.

In addition, the FSSA, in collaboration with Innovate UK, is also researching development of a a multi-sensor processing and fusion system.

Named Sapient, it enables autonomous sensors to make local decisions without constant operator attention, and fuse the output in an AI-enabled central fusion hub.

DSTL hopes that Sapient will be able to take the load off the soldier and reduces the risk of human error, as well as reducing the number of operators.

Sapient was used in the multinational experiment Contested Urban Environment (pictured) in Montreal, Canada in 2018.

DSTL is looking into multi-sensor fusion and enhanced and automated processing of sensor output to ‘keep ahead of the curve’ as sensors miniaturise and processing power increases, as well as coping with the degradation in sensor performance caused by the real environment and congested and cluttered areas of operation, explained the DSTL spokesperson.

The research is hoping to meet the growing requirement from militaries for generic sensor capabilities that are equally effective given any data feed, operational scenario, or threat.

Back to News

Share to