Improved Robot Localization
The art of determining where a robot is located through the use of sensor data and previous knowledge of the environment is know as Localization. For all of the recent technologies that are incorporated into mobile robots, most use a localization method known as Monte Carlo Localization which has barely changed in the past 15 years. In fact, a good portion of the robots in the world are still localizing like it’s 1999, because their localization software was in fact written in 1999.
During his second internship at Fetch Robotics, Luc Betteiab worked on a modular, plugin-based localization framework, which has allowed our engineers to more easily advance the state of the art in localization technology. Since Luc completed his internship last summer our navigation team has used his framework to completely overhaul our localization software, resulting in huge improvements in accuracy and robustness.
Luc completed his undergraduate degree in electrical engineering with a focus in artificial intelligence and robotics at Case Western Reserve University in Cleveland, OH. He is now finishing his master’s degree in electrical engineering before joining as a full time employee later this summer!