Author: James Male -
A soundbite by James Male, a PhD student in the Department of Electronic & Electrical Engineering.
James works in the Interaction Research Lab, which focuses on the development of human-robot collaborative technologies. His own research centres around developing methods for robots to work effectively with humans in an industrial manufacturing setting.
He hopes that this will allow robots to take over dull and dangerous jobs, allowing humans to perform complex tasks with reduced physical and mental strain.
Transcript:
"My name is James Male and I'm a PhD student in the Electronic and Electrical Engineering department. I work in the Interaction Research Lab which is focused around the development of human robot collaborative technologies. My research is primarily focused around developing methods for humans and robots to work together effectively in an industrial manufacturing setting.
"Current mass production techniques tend to use a large, powerful variables which must be caged off behind protective barriers for safety. People are then required for any complex or dextrous activity along with task requiring, thinking or adaptation. Relatively recently, collaborative robots or so-called, "cobots”, have started to make their way into industry. These operate at lower speeds and have force and torque sensing their joints, allowing them to know if they hit something so they can stop. This makes them safer to work around humans without being caged off, but they still lack any real intelligence. They predominantly just perform the same tasks again and again.
"Within the Interaction lab, there are various projects on developing sensing perception and actuation methods for human or robot collaboration. My research is looking into increasing the level of perception the robot has, so it can make decisions on how best to help a human worker. For this, I am using various action recognition, person tracking and visual recognition methods to sense the environment around the robot. A key point here is that robots currently tend to live in the present. They have very limited knowledge of the past which means they can't plan for the future, only react to whatever information they have in a specific moment. I'm therefore building a memory of past events the robot has perceived from the various sensing methods used. From this, it can learn from experience to better predict what actions the user will require it to do next.
"It’s hoped the versatile interacting methods developed will allow robots to take over the dull, dangerous and dirty jobs, allowing humans to perform dexterous and complex tasks with reduced physical and mental strain. Human-robot collaboration is one of the key technologies being developed in pursuit of industry 4.0, the fourth industrial revolution after those of mechanisation, mass production and digitisation. Key to industry 4.0 is developing smart manufacturing methods and cyber physical systems, where increased monitoring and control allows processes to become more customisable, more efficient and responding quickly to change. Perhaps a keyway to think of it is moving from automatic to autonomous manufacturing processes. Processes move from being highly specialised for specific tasks and move to worlds being highly versatile with increased intelligence and control."
Respond
