Localization and Mapping for Service Robotics Applications in Indoor Environments

Service robotics include all the robotic systems that are built to perform tasks in place of human users or in collaboration with them. Professional service robots typically augment people for execution of tasks in the workplace. Personal service robots, on the other hand, can assist people in their daily lives in their homes, or for compensation for mental and physical limitations. Despite the advancements of robotics in the automation industry, enabling robust and safe autonomous navigation in dynamic unstructured environments will be necessary for service robots to be part of our everyday life.

Robust localization and mapping algorithms were developed for long-term indoor navigation in industrial and domestic GPS-denied environments, using traditional sensors and vision sensors. A multi-robot cooperative localization algorithm was developed for the navigation of teams of mobile robots in large symmetric logistic spaces. Dynamic changes in the map are recognized and spread among team members. Then a graph-based SLAM algorithm was developed, composed of a fast back-end and a series of front-ends for different sensors modalities. The algorithms were used in a number of real-world robotics applications, partly exploiting the emerging paradigm of Cloud Robotics, which considers robots as agents connected to a remote network infrastructure; this allows them to benefit from off-board computational and storage resources and to leverage common knowledge.