At a busy recycling facility in Rainham, east London, dust hangs in the air while conveyor belts rattle through piles of mixed waste. The site, operated by the family-run Sharp Group, processes up to 280,000 tonnes of material each year, ranging from discarded shoes and broken electronics to chunks of concrete.
The environment is noisy, fast-moving, and physically demanding. Around two dozen agency workers monitor the conveyor lines, picking items as they pass at speed. The company rotates staff every 20 minutes to reduce fatigue, with brief pauses in the system to give workers short breaks.
Despite safety efforts, the recycling sector remains one of the more hazardous industries in the UK. Workplace injury and illness rates are significantly higher than the national average, and the sector’s fatality rate remains a concern. Staff turnover at the Rainham facility reaches around 40 percent annually, reflecting the difficulty of retaining workers in such conditions.
Line supervisor Ken Dordoy described the challenge of maintaining performance on the fast-moving belts, noting that some workers struggle to keep pace with the constant flow of material.
In response to these pressures, the facility has begun testing robotic assistance on the sorting line. A humanoid machine, named Alpha, has been introduced as part of a pilot programme to explore automation in waste sorting. The robot is being developed by TeknTrash Robotics using hardware from RealMan Robotics.
Unlike conventional recycling machines, Alpha is designed to mimic human movement, allowing it to operate within existing infrastructure without major redesigns. The system is still in training, with engineers guiding its movements while a human worker demonstrates sorting tasks through a virtual reality setup.
According to TeknTrash founder and chief executive Al Costa, the robot is not yet ready for full deployment. It is being trained using large volumes of visual data collected from multiple cameras that track both successful and unsuccessful picks.
The aim is to teach the system how to identify objects on fast-moving belts and physically handle them with accuracy. Costa said early development often takes time, as the system must learn from extensive real-world examples before becoming operational.
At Sharp Group, finance director Chelsea Sharp said the long-term goal is to integrate robots into daily operations, allowing machines to handle repetitive picking tasks. She added that automation could enable continuous operation without fatigue, absence, or safety risks.
Other companies are pursuing different approaches. AMP Robotics uses artificial intelligence and air-jet systems to sort materials at high speed, while Glacier deploys robotic arms trained on large datasets to handle unpredictable waste streams.
Glacier co-founder Rebecca Hu-Thrams noted that recycling facilities often encounter unexpected and hazardous items, making automation both challenging and necessary.
Experts say the shift toward robotics is becoming unavoidable. Professor Marian Chertow of Yale University said AI-driven systems combined with robotics could improve efficiency, safety, and material recovery across the sector.
Back in Rainham, however, the human reality remains visible. Workers continue to operate in loud, dusty conditions where fatigue and turnover are constant issues. While automation is still in early stages, the direction of travel is clear: machines are moving closer to the conveyor belt, and the role of human pickers may soon change significantly.
