Zheng Linjiang

RFID-based production operation management for multi-varieties and small-batch production

The rapid changing market is making manufacturing systems increasingly trend to multi-varieties and small-batch production. It drives that strict production operation management in detail is done. In this paper RFID tags are introduced to overcome difficulty to identity important production elements in workshop. A tag data model of production element is proposed, and the architecture and integration operation mode of corresponding production operation management system are established. Then, an information fusion model of real-time heterogeneous production data from multi-sources and implementation methods of rapidly reconfigurable production operation management system based on component are presented. The application results in a motorcycle assembly line show the production efficiency and quality are obviously improved.

Simulation of pedestrian behaviors based on an improved floor field cellular automaton model under attractive incidents

The study of pedestrian dynamics is very significant for promoting pedestrian traffic management and optimizing the layout of public facility in comprehensive transportation hubs, markets, theaters etc. In order to simulate and reproduce pedestrian behaviors under attractive incidents, an improved floor field model is proposed in this paper. This improved model is composed of static floor field, dynamic floor field and the new proposed attractive field, which takes into account the attractive degree of incident, view distance of onlooker, and pedestrians safe psychological distance. The transition probability of pedestrian is decided by the interaction of static floor field, dynamic floor field and attractive field. Besides, pedestrian movement state is divided into three types under attractive incidents: walking, staying, and leaving, which can depict the pedestrian behaviors of gathering, looking and dissipation. The experiment results show that our proposed model can reproduce the onlooking processes and capture the typical characteristics of torus-shaped phenomenon under attractive incidents.

A data processing model for improving RFID application reliability in logistics tracking system

The characteristics of natural radio communications and the disturbances of application environment inevitably lead to the unreliable phenomena of duplicated readings, false positive reading and false negative reading when RFID is applied to logistics tracking system, and numerous unreliable data are generated. To solve this problem, a series of evaluation indexes for evaluating RFID application reliability were put forward, and a layered data processing model was established to improve RFID application reliability, including primitive event processing layer and complex event processing layer. The former receives primitive reading events from the reader-network, cleans redundant events, and thus tidy logic reading events are obtained. And the complex event processing layer detects and corrects false positive reading and false negative reading through setting up application integrity constraint. The results of application show that the model can guarantee the correctness and integrality of the data, and improve RFID application reliability.