UNIVERSITY OF WATERLOO · INTEGRATED ROBOTICS SYSTEM

Summit XLS Multi-Floor
Autonomous Navigation

Collaborative system-level integration for multi-floor mobile-robot navigation using Summit XLS, ROS navigation, vision-based pose estimation, WAM manipulation, elevator interaction, and floor-specific maps.

Project Overview

Multi-floor navigation as an integrated robotics task

The collaborative system was developed around a practical scenario: navigate from a laboratory on the third floor to the elevator, interact with the elevator controls, transition between floors, and continue navigation toward the first-floor RoboHub.

Summit XLS Mobile robot platform
ROS Navigation & task integration
ZED + ArUco Vision-based pose estimation
WAM 7-DOF Elevator interaction
Navigation Scenario

Third-floor laboratory to first-floor RoboHub

The scenario combined floor-specific navigation, elevator interaction, manipulation, perception, and task sequencing within one ROS-based system.

01 Navigate from Lab
→
02 Reach Elevator
→
03 Call Elevator
→
04 Elevator Interaction
→
05 Board Elevator
→
06 Continue to RoboHub
My Contribution

Summit XLS autonomous navigation

My primary responsibility was the mobile-navigation subsystem using the Summit XLS platform. I worked on floor mapping, localization, ROS-based navigation, goal execution, and physical navigation testing within the multi-floor workflow.

The WAM manipulation and elevator-button interaction subsystem was developed by other members of the research team and integrated with the mobile-navigation system as part of the collaborative robotics demonstration.

Navigation responsibilities

  • First- and third-floor navigation mapping
  • Summit XLS localization setup
  • ROS move_base navigation
  • Goal-based autonomous navigation
  • Navigation from the laboratory to the elevator area
  • Navigation workflow integration
  • Physical testing and troubleshooting
Mapping & Navigation

Floor-specific maps and autonomous navigation

Navigation maps were created for both the first and third floors. The Summit XLS was configured and tested to localize and navigate within the mapped environments.

Physical navigation tests included autonomous travel from the third-floor laboratory to the elevator area as part of the integrated multi-floor navigation workflow.

Navigation work

  • First-floor navigation map
  • Third-floor navigation map
  • Summit XLS localization setup
  • ROS move_base integration
  • Goal-based autonomous navigation
  • Navigation to the elevator area
  • Navigation-system testing
System Architecture

Navigation, perception, manipulation, and task control

The collaborative system connected the Summit XLS navigation stack, WAM manipulator, ZED camera, task-control software, inverse kinematics, and networked computers.

MOBILE NAVIGATION

Summit XLS

ROS navigation and move_base were used for localization, goal execution, and autonomous movement within the mapped environments.

PERCEPTION

ZED + ArUco

Vision-based marker detection was used by the elevator interaction subsystem to estimate the pose required for button pressing.

MANIPULATION

WAM 7-DOF

The WAM manipulator subsystem used inverse kinematics and pose calibration for physical interaction with the elevator controls.

TASK CONTROL

ROS task orchestration

A task-control layer coordinated navigation, perception, manipulation, and elevator-interaction stages across the integrated system.

Collaborative Subsystem

Vision-guided elevator interaction

The elevator-interaction subsystem was developed by other members of the research team. It combined perception and manipulation within a consistent spatial reference, using ArUco markers as visual landmarks for estimating the target pose.

The WAM manipulator used the estimated pose and an inverse-kinematics solution to approach and press the elevator control. This subsystem was integrated with the Summit XLS navigation workflow.

Interaction pipeline

  • Detect ArUco marker
  • Estimate target pose
  • Transform pose into robot/manipulator frame
  • Solve WAM inverse kinematics
  • Execute button-press motion
Calibration & Debugging

Coordinate-frame alignment

Reliable system integration required consistent camera, robot, and manipulator coordinate frames.

Frame-alignment issue

System testing indicated that the camera frame and the other system frames were not consistently aligned, introducing pose offsets during manipulation.

Calibration approach

Base ArUco markers were added to the Summit platform to establish a visual reference, and larger markers were used to improve pose-estimation reliability.

System Integration

Practical robotics troubleshooting

Integrating multiple robotic subsystems required debugging beyond navigation and manipulation algorithms.

Communication, networking, planner behavior, coordinate alignment, and hardware recovery all affected whether the integrated demonstration could operate reliably.

Issues addressed

  • CAN communication failure
  • Summit control loss
  • Remote-PC network disconnection
  • Router and network reconfiguration
  • Local-planner oscillation near the goal
  • Pose-estimation and frame-alignment errors
Physical Verification

Integrated testing through elevator boarding

The project was developed incrementally through physical subsystem testing and collaborative integration.

MY CONTRIBUTION · NAVIGATION

Autonomous navigation to the elevator

First- and third-floor maps were created, and the Summit XLS was physically tested for autonomous navigation from the third-floor laboratory to the elevator area.

COLLABORATIVE SUBSYSTEM · MANIPULATION

Elevator interaction and boarding

The research team's perception and WAM manipulation subsystem was physically tested for elevator-button interaction and integrated with the navigation workflow through robot boarding.

Demonstration

Elevator interaction demo

Recorded footage from the elevator-interaction stage of the collaborative multi-floor robotics project.

This footage captures the elevator-interaction stage of the integrated demonstration. The WAM manipulation subsystem shown in the video was developed by other members of the research team, while my primary contribution focused on Summit XLS autonomous navigation and its integration into the workflow. The navigation stages were not separately recorded on video.

Project Outcome

Integrated navigation through elevator boarding

During my involvement, the integrated demonstration progressed from the third-floor laboratory through autonomous Summit XLS navigation to the elevator, elevator interaction, and robot boarding.

Further development, including autonomous navigation after the elevator transition toward the first-floor RoboHub, was handed over to subsequent researchers when I left the PhD program.

My engineering focus

  • Summit XLS autonomous navigation
  • ROS move_base integration
  • Localization and goal execution
  • First- and third-floor mapping
  • Navigation to the elevator area
  • Navigation-system integration
  • Physical robot testing and troubleshooting
← Back to Waterloo Research