An excellent article on Sundance by the Knowledge Transfer Network. Original article can be seen here.
Sundance Multiprocessor Technology Ltd was founded in 1989 with a focus on design, manufacturing and testing of electronics embedded modules, based on a leading-edge multicore technology. This technology allows infinitive scaling of processing power and provides a large range of sensor interfaces and numerous instances of control functions.
With a small team of twelve people, who cumulatively possess 100s of years of experience, the focus of Sundance is still multiprocessing technology. Their key areas of interest comprise of:
- Vision System and sub-components for safety
- Advanced Sensor technology for monitoring
- Robotics electronics for control and unmanned vehicles
- Deep Learning and Edge-AI
Flemming Christensen founded Sundance, when their first products were designed for the parallel processing market. Many years later, Precision Robotics & Edge-AI has become the new focus and according to Flemming, “is the most interesting technology, in my view, since the beginning of the CPU silicon revolution in the ‘70s.”
Sundance have been involved in several key projects, more recently:
- Development of a Vision, Control & Sensor solution for Precision Robotics as part of the H2020 project, TULIPP.
- Been involved with a robotics project, VineScout, for intelligence gathering in vineyards.
- Began an Innovate UK project, ARISE, for the mining industry in December 2019.
- Involved with a Eurostars funded project called FieldCompanion that will use Edge-AI for Smart Spraying. Details have now been launched here.
To build on their expertise and take Sundance into a new future, the organisation is interested in any collaborations that will take their current precision robotics platform to the next level. Part of this involves streamlining the way Edge-AI solutions are produced for customers to integrate into their existing robotics/automation products. Their other aim is to improve on building high-tech and bespoke electronics in low-volume, but high-value for Sundance, while still being cost effective to customers.
“We are here to deliver, and we have only just started.”
Autonomous Robotic InSpEction (ARISE) Project
The modern world is dependent on the supply of mineral resources to provide the raw materials to meet, current and future needs for energy, manufacturing and construction industries and demands are increasing exponentially due to population growth, life expectancy and a more technologically-focused, digital world which requires sophisticated mineral products.
The mining industry is committed to operating safely and reducing accident numbers, and it is increasingly migrating underground as surface deposits are exhausted. The underground environment is challenging due to: high rock stress, high temperatures, poor communications with surface, restricted access and lack of access to satellite positioning systems.
The key risk in deep level mining is geotechnical because ground conditions are exacerbated by high stress at depth and the unpredictable nature of the blasting process on stress redistribution. It is essential to undertake disciplined geotechnical inspections of newly exposed rock after every blast, but it is well known that this process is dangerous, time-consuming and subject to human error.
ARISE aims to implement autonomous surveys of geotechnical conditions during the normally unproductive period immediately after the blast when workers vacate the mine due to post-blasting fumes and seismic risk.
The robotic platform will be used to:
- Survey roof conditions in newly-blasted areas
- Monitor material flow in orepasses and extraction points, particularly mapping ‘hangups’ that can block orepasses. Mapping hangups from below is extremely dangerous for people
- Accurately map areas in 3D for reconciliation and design verification.
ARISE will provide safety and financial benefits while not affecting the production cycle (operating in the shift change periods) and is therefore attractive for industrial roll-out.