Mission Planning for Indoor Quadrupedal Robot Navigation Assistance for Visually Impaired People

The Robodog Flagship Project at PBL aims to enhance the standard of living for the visually impaired by developing a navigation assistance robot capable of guiding them to specific locations while avoiding obstacles. The project's next semester goal is to improve the robot's navigation system, enabling it to efficiently and safely guide a blind person in complex indoor settings, specifically moving between rooms on different floors in ETH buildings. Our existing system consists of a Unitree A1/Go1 quadrupedal robot equipped with additional sensors (depth cameras, lidars). It is currently capable of autonomously navigating flat indoor areas using SLAM or an existing map. The upcoming semester's objectives include improving the locomotion controller, improving state estimation, expanding the human-robot interaction, and enabling the system to reach any room across different floors. This project will focus on the mission planning for these navigation tasks, including entering the correct elevator, setting the goal floor, waiting until the lift reaches the required floor and leaving it. All of this while also guiding a visually impaired person. For some subtasks the Robodog offers better perception, while in others the robodog requires inputs from the guided person. Your role in this project begins as soon as the Robodog, guided by a high-level planner, approaches the elevators. The high-level planner provides a destination floor to reach where it will take over again. The tasks to reach the destination floor are numerous and must be coordinated to safely guide the person in and out of the elevator. A communication system is integrated to facilitate simple commands and feedback between the Robodog and the guided person. This is a hands-on project where you will engage directly with robotic hardware in a collaborative team environment. You'll work alongside peers focusing on different system components, with the collective goal of realizing the project's vision. Note that the voice interface and the elevator destination setting functionalities are being developed in parallel by other projects, necessitating strategic coordination for a fully functional system. **Prerequisites** - Solid programming experience (Python). - Experience with ROS or willingness to pre-learn and practice. - Motivation to work in a team environment. - Innovative engineering thinking and creativity for solving a non-trivial task **Character** - 50% Implementation - 30% Experimentation and tuning - 10% Evaluation and Analysis - 10% Report and Presentation **Thesis Type** Bachelor Thesis, Semester Thesis

The Robodog Flagship Project at PBL aims to enhance the standard of living for the visually impaired by developing a navigation assistance robot capable of guiding them to specific locations while avoiding obstacles. The project's next semester goal is to improve the robot's navigation system, enabling it to efficiently and safely guide a blind person in complex indoor settings, specifically moving between rooms on different floors in ETH buildings.

Our existing system consists of a Unitree A1/Go1 quadrupedal robot equipped with additional sensors (depth cameras, lidars). It is currently capable of autonomously navigating flat indoor areas using SLAM or an existing map. The upcoming semester's objectives include improving the locomotion controller, improving state estimation, expanding the human-robot interaction, and enabling the system to reach any room across different floors.

This project will focus on the mission planning for these navigation tasks, including entering the correct elevator, setting the goal floor, waiting until the lift reaches the required floor and leaving it. All of this while also guiding a visually impaired person. For some subtasks the Robodog offers better perception, while in others the robodog requires inputs from the guided person.

Your role in this project begins as soon as the Robodog, guided by a high-level planner, approaches the elevators. The high-level planner provides a destination floor to reach where it will take over again. The tasks to reach the destination floor are numerous and must be coordinated to safely guide the person in and out of the elevator. A communication system is integrated to facilitate simple commands and feedback between the Robodog and the guided person.

This is a hands-on project where you will engage directly with robotic hardware in a collaborative team environment. You'll work alongside peers focusing on different system components, with the collective goal of realizing the project's vision. Note that the voice interface and the elevator destination setting functionalities are being developed in parallel by other projects, necessitating strategic coordination for a fully functional system.


**Prerequisites**

- Solid programming experience (Python).
- Experience with ROS or willingness to pre-learn and practice.
- Motivation to work in a team environment.
- Innovative engineering thinking and creativity for solving a non-trivial task


**Character**

- 50% Implementation
- 30% Experimentation and tuning
- 10% Evaluation and Analysis
- 10% Report and Presentation


**Thesis Type**

Bachelor Thesis, Semester Thesis

- Familiarize with the current Robodog system - Use theRobodog's camera to choose the correct elevator - Understand and work with the current local planner to navigate the robot into the elevator. - Design and implement a state machine to manage various phases, such as elevator waiting, responding to human feedback, entering and exiting the elevator, and more. - Interact with the voice communication interface for human-robot interaction.

- Familiarize with the current Robodog system
- Use theRobodog's camera to choose the correct elevator
- Understand and work with the current local planner to navigate the robot into the elevator.
- Design and implement a state machine to manage various phases, such as elevator waiting, responding to human feedback, entering and exiting the elevator, and more.
- Interact with the voice communication interface for human-robot interaction.

Davide Plozza (davide.plozza@pbl.ee.ethz.ch) Paul Joseph (josephp@student.ethz.ch) Steven Marty (martyste@pbl.ee.ethz.ch)

Davide Plozza (davide.plozza@pbl.ee.ethz.ch)
Paul Joseph (josephp@student.ethz.ch)
Steven Marty (martyste@pbl.ee.ethz.ch)