PAL Robotics | Assuring Autonomy International Programme | ALMI Project

In this talk

1. 0:01

   Planning routes around uncertainty

   The project develops ways for Tiago to handle disruptions in a home setting, with meal preparation and emergency response as examples. The team built a Gazebo simulation and represents accessible locations as a graph, using route reliability estimates that account for connections and possible furniture obstructions. When a person is present, the planner accounts for their assumed shortest route and adjusts Tiago’s path to reduce collision risk.

2. 2:17

   Starting a cooking task

   Human recognition runs on an NVIDIA Jetson NX, while Acapela generates the robot’s spoken responses. PAL Robotics’ memory map, integrated by the University of York, tracks item locations. In a pasta scenario, Tiago asks Delia whether she wants to cook, fetches the pasta, and guides her to boil water.

3. 2:55

   Grasping and handling kitchen objects

   Tiago’s Advanced Grasping package combines vision, behavior trees, and MoveIt planning. Its object and table detection steps can be swapped to suit different tasks. OctoMap visualizes the robot’s digital twin and mapped surroundings; objects change color after a successful grasp, based on end-effector feedback. Tiago brings the pasta and tomato sauce to the user and can place objects on a chosen surface.

4. 7:41

   Responding to emergencies

   While roaming the kitchen, Tiago uses the OpenDR Toolkit to detect a person who has fallen. The robot can use Twilio to call a designated contact, and when it encounters someone, it explains that help is needed and asks them to assist.

5. 8:48

   Supporting older adults

   The project’s results inform understanding of user needs and system requirements for robots that support older adults with cognitive impairment.