Driverless cars and delivery drones promise to make life easier as well as saving us time and money. What’s not to like? The biggest concern for many is, despite assurances that AI will be far less error prone than human drivers and controllers, resulting in a huge reduction of accidents, bugs in the software of robot drivers causing horrendous pile-ups or colliding drones dropping debris upon us from a couple of hundred meters up.
Obviously, ensuring the former scenario rather than the latter turns out to be the reality is the number one priority of any commercial operation engaged in bringing mobile robo-controlled vehicles or other technology to market. As such, the latest technology in the world of AI is going hell for leather on the development of navigational and avoidance systems. The potential commercial value of the best systems is also proving to be an attraction for investors, something SLAMcore, a company spun out of Imperial College London has taken advantage of.
SLAMcore, which is working on AI technology to help robots avoid crashes has attracted a $5 million (£3.8 million) investment, with recently closed round led by venture capital firm Amadeus Capital Partners and also involving Toyota’s venture capital arm. Its advanced navigation system, call ‘simultaneous localisation and mapping’, or SLAM, is designed to help robots and drones navigate and avoid collisions in unfamiliar surroundings or dense urban environments.
The biggest commercial application of SLAM technology, variations of which many research teams around the world are working on, is driverless cars. However, pretty much every kind of robot technology that involves movement, all the way down to robo-vacuum cleaners, uses some version of the system.
SLAMcore, which was established by a group of academics from Imperial College with research led by Dr Stefan Leutenegger and Dr Jacek Zienkiewicz with Own Nicholson in the chief executive role, plan to make their system licensable. That will mean robotics companies that don’t have the resources of an Alphabet or Amazon to develop their own sophisticated navigation AI, won’t have to. As explained by Nicholson:
“We’re really trying to democratise robots. It shouldn’t be the reserve of just a handful of tech giants because that’s not good for innovation.”
SLAMcore’s technology will mean smaller companies can “get on with coming up with crazy new ideas” and use the spin-off’s system to help “bridge the gap between demos and commercially viable systems”.
The investment funding secured will mean SLAMcore can now develop detailed mapping solution that will create
“geometrically-accurate reconstructions of the robot’s surroundings in real-time and understanding the objects within, utilising the latest development in machine learning.”