A new control principle - Control by re-planning
Tomorrow’s train traffic, with higher speed, more frequent traffic, mixed traffic and many independent traffic operation companies, requires new principles and technical solutions for efficient train traffic control. Today's control systems are often designed to support the operator's possibilities to react on alarms, conflicts and disturbances and to solve acute problems and conflicts. However, in order to perform efficiently, operators should be able to follow the dynamic development of the traffic system over time and prevent disturbances. In order to achieve this, we must change the control paradigm from technical control of the infrastructure into higher level traffic planning tasks. This is done by replacing the traditional control commands by real-time re-planning.
Advanced laboratory prototypes have successively been implemented and tested. By connecting user interface prototypes to a train traffic simulator it has been possible to perform experiments with the design of new user interfaces and decision support tools, and to test and evaluate new control strategies for the train traffic control operators.
Our research has been based on a very detailed description and analysis of how train traffic is controlled today, the mental models of the dispatchers and the strategies they use for decisions and control tasks. The research has consisted of mainly the following steps:
- Observations and interviews with dispatchers and other professionals at the traffic control centres. Analysis of the findings and identification of problems and development areas.
- Seminars with experienced and responsible professionals from the national rail and traffic control administrations. Here the visions and restrictions for future development of control systems have been specified.
- Iterative specifications and evaluations with the help of a working group consisting of experienced operative traffic control professionals.
- Tests and evaluations in a laboratory control room environment using a train traffic simulator system.
In order to support real-time planning of train traffic we provide the traffic controller with an interactive computerized time-distance graph. Prototypes of new user interfaces that support the new control strategy have been designed, implemented and preliminary tested in the laboratory environment at Uppsala University. The interface is designed to integrate all decision relevant information into one unified interface and to support continuous awareness of the dynamic development of the traffic process.
The computer based time-distance graph is designed in such a way that it visually supports the operators’ situation awareness of the current status and the projection into the future. The user interface, with its planning view, can support early detection of upcoming conflicts, identify possible re-planning alternatives and their predicted effects. The new control strategy has a potential to better support the traffic controller’s ability to handle continuous re-planning, with the goal to always have a functional traffic plan at hand. This plan can be automatically executed except when technical malfunctions hinder this. Automatic functions that support execution of the traffic plan must be transparent, predictable and easy to understand. The automatic functions must never change the controllers’ traffic plan but are only allowed to strictly execute the actual traffic plan. The traffic plan mainly consists of time table and track usage information, including maintenance work. Detailed interface design, easy to interpret, concerning the automation helps to keep the human in-the-loop and to avoid automation surprises. By re-planning, the operator is in control of what the automatic function will do and when. Hence, the operator is continuously in full and active control.
We have also evaluated different approaches to include decision support systems in operative train traffic control. We have found that more advanced automated decision support systems are today not a realistic alternative of several reasons. More research and development of methods are needed in this field. We have decided to focus our efforts on supporting the controllers through better presentation of information, improved information observability and quality, help with early detection of conflicts and disturbances, identification of possibilities and limitations for re-planning and evaluation of effects of alternative actions.