Automated Transport and Retrieval System (ATRS)

Automated Transport and Retrieval System (ATRS)

Project Update

This project concluded at Lehigh in August 2010. ATRS received FDA approval and entered the commercial market in June 2008.

System Overview

The Automated Transport and Retrieval System (ATRS) represents a new mobility paradigm for drivers in wheelchairs. By leveraging robotics and automation technologies, ATRS allows independent mobility without permanent modifications to the vehicle. Users can even lease vehicles, and have the ATRS components removed and reinstalled in another when the lease expires. This is in stark contrast to traditional “van conversion” solutions, which require significant permanent alterations to the vehicle. ATRS also allows the driver to be seated in a traditional vehicle seat system rather than the wheelchair itself, making it a safer transportation solution. Lastly, by avoiding the dramatic and expensive modifications required in vehicle conversions, ATRS will cost less. The production version of “ATRS Lite” (below right) was unveiled at the annual NMEDA conference in February 2008.


ATRS can be decomposed into five primary components: a “smart” power wheelchair system, a LIDAR system for localization, a powered lift platform, a traversing power seat, and a touch-screen user interface (UI) computer. The smart wheelchair and localization systems are the heart of ATRS. Combined, these allow the operator to be separated from the chair and eliminate the need for an attendant.

Referring to the above right figure (ATRS concept diagram), when the operator returns to his/her automobile, a keyless entry is used to both unlock the vehicle and to deploy the traversing driver’s seat. The operator then positions the wheelchair, and performs a seat-to-seat transfer (pose A). After this, the wheelchair is deployed to the rear of the vehicle (pose B). Once the chair enters the LIDAR’s field-of-view at the rear of the vehicle (pose C), it is automatically tracked. This enables the van-side computer to transmit real-time control inputs to the chair over a dedicated RF link for reliable docking (locking in place) onto the lift platform (pose D). With the chair docked the operator actuates the lift via the UI – stowing the platform and chair into the vehicle cargo area. The process is repeated in reverse when disembarking.

Wheelchair Localization

To reliably execute docking under a broad range of environmental conditions, wheelchair localization requires both robust feature segmentation as well as accurate pose estimation with respect to the lift platform. The platform mounted LMS-291 LIDAR system provides bearing, range and reflectivity measurements that are leveraged for robust feature segmentation. The positions of these features – in conjunction with the control inputs to the chair – are then used as input to an Extended Kalman Filter that estimates the wheelchair pose over time.


Component Level Testing

Extensive testing was done at the component level. For example, to characterize localization system performance, turntable testing (left figure) was conducted. This allowed simulated wheelchair position estimates to be projected to a single point (the center of rotation) so that “ground truth” localization accuracy could be established. The rotation speeds also closely mimicked the maximum expected wheelchair fiducial velocities.

Sample data collected at the 2 meter range, along with mean absolute errors are shown below. At all ranges of interest, the system exhibited sub-centimeter localization precision. Additional subtests investigated the effects of damage to the retroreflectors, heavy rain, and interference from secondary light sources on localization performance. Additional information can be found here — paper (Google Scholar) and poster

Screen Shot 2017-06-21 at 8.52.36 PM

System Demonstration

The beta system was demonstrated over 3 days at the World Congress Exposition on Disabilities(WCD2006)– video //vader

Screen Shot 2017-06-21 at 8.52.47 PM

Previous Work: Vision-based Control

In order to reliably stow and deploy the wheelchair to/from the lift platform autonomously, the navigation task is decomposed into two phases. Navigation from the operator position to the vicinity of the lift platform is accomplished using laser based feedback control with sensors onboard the chair itself. In contrast, docking and undocking the chair on the lift platform relies upon a computer vision system on the minivan to estimate chair pose. This is motivated by the tight docking requirements.

The lift platform/docking station is designed to accommodate a range of power chair models. When used with wider wheelbase models, the available clearance for navigating the chair onto the lift platform is approximately 2-4 cm on each side. This drives the need for highly accurate estimates of chair position and orientation relative to the lift platform. Significant noise in the measurements makes this level of accuracy difficult using a laser-based localization approach. Instead, we employ a high-resolution 1024×768 DragonflyTM camera mounted within the van’s liftgate (see illustration above) to track the chair’s pose at 15 Hz. After a proper calibration, the vision system can localize the chair to a tolerance of < 1 cm.

Screen Shot 2017-06-21 at 8.53.08 PM

Videos (vader.cse)

  1. ATRS Outdoor Trials (Nov 2007) [ wmv, 14.12 MB ]
  2. Commercial Video From Freedom Sciences
  3. ATRS Demonstrated at WCD (Nov 2006) [ xvid, 31.3 MB ]
  4. ATRS Proof-of-concept Demonstration (June 2005)
  5. Docking Experiments
  6. ATRS Concept Animation


  • C. Gao, I. Hoffman, T. Miller, T. Panzarella, and J. Spletzer, “Autonomous Docking of a Smart Wheelchair for the Automated Transport and Retrieval System (ATRS)”, Journal of Field Robotics, Volume 25, No. 4-5, Pages 203-222, April 2008.
    Full Text (PDF) – Google Scholar.
  • C. Gao, I. Hoffman, T. Miller, T. Panzarella, and J. Spletzer, “Performance Characterization of LIDAR Based Localization for Docking a Smart Wheelchair System,” in IEEE/RSJ International Conference on Intelligent Robots and Systems (IROS 2007) Workshop on Assistive Technologies: Rehabilitation and Assistive Robotics, October 2007
    Full Text (PDF) – Google Scholar – poster (PPT)
  • C. Gao, I. Hoffman, T. Panzarella, and J. Spletzer, “Automated Transport and Retrieval System (ATRS): A Technology Solution to Automobility for Wheelchair Users,” in the 6th International Conference on Field and Service Robotics (FSR 2007), April 2007
    Full Text (PDF) – Google Scholar.
  • H. Sermeno-Villalta and J. Spletzer, “Vision-based Control of a Smart Wheelchair for the Automated Transport and Retrieval System,” in the 2006 IEEE International Conference on Robotics and Automation.
    Full Text (PDF) – Google Scholar.
  • Edmund F. LoPresti, Jim Osborn, Thomas Panzarella, Thomas Panzarella, Jr., and John Spletzer, “Automatic Transport and Retrieval System for Power Wheelchairs,” RESNA, June 2005.

Media Coverage



  • Lehigh University
  • Freedom Sciences
    • Tom Panzarella
  • Carnegie Mellon University
    • Sanjiv Singh
    • Jim Osborne