deliverable 2.3 : Demonstrators warning system
author : Thorsten Szczepanski
The main aim of the Prospie system is the protection of a worker in hazardous industrial areas. Today a worker usually decides on an individual basis about his/her current state of health as well as his ability to remain under the environmental conditions. These subjective decisions often lead to a wrong assumption about the possible work time and can cause health problems to the worker. Therefore a decision mechanism based on objective criteria is needed. To achieve this a near-body automation system was envisioned in the project. As any such system it consists of the three parts sensor, controller and actuator as well as a dedicated communication system. In order to allow a remote operator to monitor critical situations remotely a supervisory software system must also be provided.
The sensors of the systems are divided into body sensors, inner-garment sensors and environmental sensors. An extensive description of the used sensor types can be found in deliverable D2.1 and 2.2. All sensors are connected to a pre-processing and data-transmission unit attached to each PPE, the so called gateway. The gateway allows the implementation of medium sized control algorithms for the warning system. The output of the control algorithms decides the actions of the warning system.
The control algorithms to be implemented can be divided into 3 categories. The first category comprises simple decision algorithms. It is based on event processing mechanisms. An event is defined to be the violation of a trigger value by a sensor value. These thresholds can be set according to the experience of a human operator. The second class of algorithms is based on regression analysis of experimental data. It provides a list of parameter dependencies from which more elaborate decision algorithms can be derived. The third class of algorithms utilizes a complex biological model of the human physique in the context of environmental conditions. During the project the model will be validated against the data used in the regression analysis. Only the algorithms of the first and second class can be implemented directly on the gateway. The model-based algorithms most certainly will be included into the remote supervisory system.
The warning system of the Prospie system consists of the component warning actuators on the PPE and a status and event-Monitoring as part of the supervisory system. The warning actuators are intended to be placed on the outside of the PPE, on the so called layer 3 of the Prospie clothing model. As discussed in Deliverable D2.2, the layer consists of three multi-purpose communication ports and power sockets. They allow for the attachment of any I2C based sensor or actuator (as long as they adhere to the mechanical restrictions). The connection between the sensor/actuator is provided by snap-fasteners. These so called patches can be used to design a bespoke PPE depending on the needs of a customer. As exemplary warning patches an optical and an acoustical warning patch were developed.
The supervisory software of the demonstrator was extended to provide for the Event-processing algorithms. The events which form the foundation of the decision algorithm can be defined by the operator. Every available sensor value can be used in the event definition. The events may be weighed according to their relative frequency. During run-time the resulting event-list is stored in an event-log. Depending on the occurring event combinations the status of the monitored PPE is assessed and displayed in the software. If a warning patch is attached to the PPE the operator may as well send a status information to the patch in order to inform the worker about a critical condition. This mechanism is only intended as a demonstration of the working condition of the system. In future the algorithms of the above discussed classes 2 and 3 shall be used. The class 1 algorithms can only be seen as a fallback strategy which supersedes all local algorithms and allows the operator to interfere if necessary.
For an overview of the system the reader is referred to the Prospie Demonstrator video, which forms the scientific part of this deliverable. The link of this video is http://www.youtube.com/watch?v=syBXs5laEic