Routine use of robotic lift trucks may not be far off.
Researchers at MIT's Computer Science and Artificial Intelligence Laboratory, along with a team of engineers, have transformed a Toyota 8-Series lift truck into an autonomous bot capable of working alongside human supervisors using voice commands or hand gestures.
The 3,000-pound-capacity lift truck from Toyota Material Handling (TMHU) is capable of locating, lifting, moving, and placing supplies while traversing just about any type of terrain. It was demonstrated last month at an event hosted by the U.S. Army Logistics Innovation Agency at Fort Lee, in Virginia.
"Robotic forklifts have the potential to protect both military and civilian personnel working in high-risk environments, such as hazardous material storage facilities, said Brett Wood, president of TMHU.
The demo included a review of the robot's safety features, sensor capabilities, and human-robot interface.
The researchers and engineers developed a complex network of systems to enable the lift truck to navigate real-time conditions faced by lift truck operators (navigating obstacles and interacting with other moving vehicles, for example). To do this, they added a camera, sensors, laptops, servomotors, Wi-Fi, and a PDA.
"We chose the internal combustion Toyota lift truck because it can be operated outdoors on packed earth or gravel and because, with mini-lever control some of its functionality can be controlled electronically rather than solely mechanically," said MIT Professor Seth Teller, who headed the project.
The modified vehicle wirelessly exports video from its own point of view, so the human supervisor, even if hundreds of miles away, can see whatever is nearby (provided there is network connectivity between the lift truck and supervisor's tablet).
In September 2009, for example, the team demonstrated the lift truck operating autonomously at MIT, in Cambridge, Mass., while under the supervision of an operator in Washington D.C.
The lift truck uses sensors and perception algorithms to form an internal representation of its surroundings; a human-robot interface to determine what task the operator wants completed; and planning algorithms to plan a set of vehicle motions and fork actions to satisfy the task, engineers explained. The fork lift then autonomously executes the plan while algorithms monitor for unexpected situations so the scheme can be changed on the fly. The robot also listens for shouted human speech, and will pause if it concludes that a human is yelling a warning.
The lift truck's system was designed to operate outdoors on uneven terrain such as gravel, and alongside people on foot and other vehicles that may block or cross its path.
Over the course of the two-year project, MIT worked with BAE Systems and Lincoln Laboratory in collaboration with the LIA; the Combined Arms Support Command Sustainment Battle Lab; the Office of the Secretary of Defense; and the Director of Defense Research and Engineering.