Rehabilitation Robotics Jumpstation
Thesis - HF Machiel Van der Loos
A history list design methodology for interactive robotic systems (1992)
In interactive computer controlled systems, history lists store records of user and system events. These lists provide objective accounts of system operation to users, analysts, designers, management and regulatory agents. History lists have proven their value with aircraft flight recorders, chemical and power plant process control, and software usability testing. These areas are distinguished by the high social cost of system malfunction and the high benefit of performance gains. For the emerging class of semi-automated health care systems such as human-service robots in rehabilitation, surgery and patient care, well-designed history lists can likewise help ensure safe and effective system operation. In this domain, however, the potential of history lists is still largely unexplored.
This dissertation presents a methodology for the systematic design of history lists for interactive robots and other computer-controlled systems. The proposed design framework consists of five guidelines:
- Identify the type of client who will use the history list information.
- Identify the purpose for which the client will use the information.
- Define the data acquisition system.
- Define the structure and content of the data.
- Determine the limits of history list applicability.
The guidelines are derived from five years of experimental history list usage. The history lists were recorded by a data acquisition system originally designed into the interface of DeVAR, a voice-controlled desktop robot for individuals with severe physical disabilities. This dissertation examines 3000 hours of history list data and 30 hours of videotaped DeVAR operation. Analyses of the history lists (e.g., regression, state transition networks, task action grammar, maximal repeating pattern) illustrate how the client's needs dictate data structure and system implementation. Tests of the methodology with DeVAR also show that: (1) modifications to the interface could increase system safety and effectiveness, (2) modifications to the history list structure could facilitate analysis, and (3) extensions to the history lists could expand the types of users and uses through close alignment with each of the five guidelines.