Danu Robotics is developing a revolutionary robotic waste sorting system to help the waste management industry significantly increase their productivity, prevent valuable resources going into landfill, boost the circular economy and clean up the environment.

Background

The world generates 3 billion tonnes of domestic solid waste annually, less than 10% of it is recycled. Worldwide, recycling sites require human intervention to pick out contaminants, which can pose health risks and is extremely inefficient.

Xiaoyan Ma founder of Edinburgh-based, clean tech company Danu Robotics, has combined her expertise in high performance computing (HPC) with her passion for the environment to revolutionise the efficiency of the recycling sector by developing an AI-powered, robotic sorting system.

She explained: “I have been a committed environmentalist since I was a teenager and always recycled my household waste, but I’d never thought about where it ended up. So, a couple of years ago, while I was studying, I decided to look into the whole process, and I was shocked at how inefficient it was.”

The Challenge

As a team of one, Xiaoyan needed help in identifying additional resources to support the development of the robotic solution. She required both experts in robotics and hardware development. She also called for help with software development and computer vision expertise. Experts in advanced data analytics and image recognition capability, would help develop a more accurate machine learning algorithm and object classification to enable the robotic system to differentiate between recyclable materials and general waste.

Following an introduction from John Hill, her student enterprise advisor at Edinburgh Innovations, Interface made several connections to different academic teams to support Danu Robotics on their journey of development.

The Solution

Interface connected Danu Robotics into the Design Manufacturing & Engineering Management (DMEM) department, at the University of Strathclyde. The DMEM students undertook a six-month project researching and developing the robotic picker equipped with a camera to identify objects and an appropriate algorithm to instruct the picker to pick out the recyclables and place them in a designated area. The robotic picker needed to meet predefined performance requirements, and in addition, the solution had to be durable, recyclable/sustainable, affordable, portable and re-programmable. Several options were considered to design robotic pickers that were fit for purpose ranging from, a custom solution where the robotic picker uses a suction and release mechanism rather than grab and release, to the modification of an off-the-shelf robot arm/picker to the combination of hard robot and soft robot.

In conjunction to this project, Interface also connected Danu Robotics into EPCC at the University of Edinburgh to develop software for identification and classification of objects and to define suitable hardware, including sensors and cameras, for the robotic picker.

The software development phase required development of a machine learning algorithm that takes image data and sensor data to differentiate recyclables from the general waste. The company had to build up a waste image database to help the system identify contaminants, the collection of the image data was supported by Glasgow City Council. Each item in this visual database was then labelled by a specialist data processing company and the updated database used to ‘train’ the machine learning algorithm to identify what can and cannot be recycled.

With initial system training complete, the software required further development to direct the robotic sorting system to remove contaminants from a moving conveyor belt as efficiently and effectively as possible. Working with EPCC’s Cirrus supercomputer resources, accelerated the development of the project, with two months of lab tests to integrate the software with the robotic hardware, followed by a three-month trial of the prototype system at Glasgow City Council’s recycling centre.

The initial collaboration with EPCC was funded by a SFC Innovation Voucher, then EU Horizon 2020 (H2020) funding which in turn helped leverage £70K in a SMART Scotland grant. They have recently secured SFC Advanced Innovation Voucher funding to continue development work with EPCC.

Other opportunities which Interface have been instrumental in assisting Danu Robotics with include:

• An MSc project for a design engineering student from Heriot Watt University interested in robotic design and kinematics to design a robotic arm that can sort waste items quickly and lift loads of up to 20kg while the entire arm assembly is moving at high speeds. An optimal design needed to consider the harsh dirty and dusty operating environment and the high speeds that the assembly needed to work in. The project provided an opportunity to produce an innovative and challenging piece of academic research, but also engagement in industrial research and product development.
• Another MSc opportunity with DMEM students at the University of Strathclyde involved the development of a control system for the waste sorting robot. Students interested in control systems, kinematics, and industrial design worked to identify a closed loop control system best suited for application in a heavy industrial robotic waste sorting system.
• A further project is underway with Design Engineering students at Heriot Watt University looking at the design and development of prototype of a new emergency stop system which needs to be designed to bring the mass to a stop in a safe manner. The robotic system uses a belt drive system driven by an industrial servo motor with no service break and currently relies on the frictional losses present to come to a stop in case of an emergency or loss of motor torque.

The Benefits

• In September 2023, Danu Robotics joined the prestigious Octopus Ventures Springboard programme and CEO Xiaoyan Ma gave a keynote at the Carbon13 conference.