Just as motorists observe road rules, most pedestrians follow etiquette when walking through a crowd (e.g. stay right, pass on the left, keep your distance.
Engineers at MIT have recently created an autonomous robot with “socially aware navigation,” that can keep pace with foot traffic while observing codes of pedestrian conduct.
In drive tests performed inside MIT’s Stata Center, the robot successfully avoided collisions while keeping up with pedestrians. The researchers detailed their design in a paper that they’ll present at the IEEE Conference on Intelligent Robots and Systems in September.
For a robot to make its way autonomously through a heavily trafficked environment, it must solve four main challenges:
- Localization (knowing where it is in the world)
- Perception (recognizing its surroundings)
- Motion planning (identifying the best path to a destination)
- Control (physically executing its desired path).
The researchers outfitted the robot with off-the-shelf sensors (e.g. webcams, a depth sensor, and a high-resolution lidar sensor). For localization, they used open-source algorithms to map the robot’s environment and determine its position.
Other studies used “reactive-based” approaches. In such research, robots are programmed using geometry or physics to calculate a path that avoids collisions.
The problem here is that human nature is unpredictable. People hardly ever walk a straight way, and instead randomly veer off. In such an environment, such robots tend to run into people.
The team found a way around such limitations by programming the robot to adapt to unpredictable pedestrian behavior as it moves through the crowd while following social codes of conduct.
They used reinforcement learning, where they did computer simulations to train a robot to take certain paths. The robot’s approach is based on the speed and trajectory of other objects in the environment.
The team incorporated social norms into this training phase as well. They encouraged the robot in simulations to pass on the right and penalized the robot when it passed on the left.
Here’s a short video that summarizes the research.