Robotic Wheelchair Takes Elderly Customers Shopping

[Kyoto, Japan] On March 30, 2011, researchers at ATR Intelligent Robotics and Communication Laboratories demonstrated the latest developments in the Ubiquitous Network Robot project, this time featuring a robotic wheelchair in a shopping mall. The demonstration featured autonomous planning and safety using ubiquitous sensor networks, location-based services, and integration with remote teleoperators and mobile devices over the internet, giving elderly wheelchair users the independence to go shopping on their own.

System Overview
The concept behind the service is that elderly people with disabilities tend not to leave home by themselves because they and their families are concerned about possible accidents. The system presented in this demonstration helps allay these fears and ensure safety using the Ubiquitous Network Robot (UNR) Platform, a networked robot system which connects robots with environmental sensor networks, remote operators, planning and processing servers, and mobile devices.

Shopping Mall Demonstration
The demonstration began as an elderly customer connected to the system using her smartphone and reserved a wheelchair robot for a shopping trip. Based on the phone's GPS coordinates, the system was able to estimate her arrival time and have a robotic wheelchair waiting for her at the entrance to the shopping mall. When the customer arrived, she got into the robotic wheelchair, which she could then drive freely using a Nintendo Wii controller.

Safety Systems
The UNR Platform helps provide safety in two ways. For visible objects such as walls and benches, on-board laser rangefinders are used to prevent collisions, a common approach used in robotics. However, a more difficult problem is that of "invisible obstacles" such as descending stairs, automatic doors, or streets where there may be traffic. These areas cannot be detected by sensors, but can pose a risk to the customer. To provide safety in these dangerous areas, a remote operator provides navigation support to the customer.

Using a map information server, the robotic wheelchair automatically stops and calls an operator when it enters one of these dangerous areas. The UNR Platform then connects the robot to an expert operator who uses a 3D graphical interface incorporating rich map data, human position data, and video from an omnidirectional camera to safely drive the customer through or around the dangerous area.

Other Services
In addition to the safety mechanisms, several other services can be provided using the robotic wheelchair and UNR system. In this demonstration a network of laser rangefinders was set up throughout the shopping mall to anonymously track the trajectories of customers. This trajectory information was used for additional safety, as well as for dynamic path planning through crowded areas. This system enables autonomous driving support for the customer in places where navigation is difficult, such as near automatic doors, and when the customer is tired. The system also helps to recommend and navigate to destinations in the mall based on the items the customer usually buys.

Results
The system thus successfully enabled the customer to move freely throughout the shopping mall by herself, giving a new level of independence to someone who typically would be dependent on a caretaker to accompany her if she even went out at all. Technologically, the demonstration presented new applications of networked robot systems, demonstrating seamless integration between sensor networks, autonomous planning, manual teleoperation, and shared autonomy to ensure safety and ease-of-use of the robotic wheelchair.