Yining Sun

SPORTS BIOMECHANICAL INFORMATION ACQUISITION AND EVALUATION FOR KAYAKING EVENTS

Intensive understanding of sports biomechanical principles is an essential issue for sprint kayaking. In recent years, the authors have studied the acquisition of biomechanical information and the evaluation of competitive ability on sprint kayaking. In this study, first, an oar force sensor was developed to detect the stroke forces applied by paddler. The sensor was attached to a paddle with minimum mechanical involvement, hence the mechanical properties of the paddle was not changed. Second, an on-water instrument of biomechanical information acquisition was developed to detect stroke force, stroke frequency, stroke power, kayak speed, and acceleration synchronously. Third, 12 evaluation factors for kayaking ability were proposed based on the factor analysis. The 12 factors include (1) the stroke rate; (2) the speed fluctuation; (3 and 4) the hull pose (L/R); (5 and 6) the stroke force (L/R); (7 and 8) the endurance (L/R); (9 and 10) the stroke technique (L/R); (11 and 12) and the stroke rhythm (L/R). Finally, the stroke rate and the forward stoke technique was analyzed briefly.

Modeling and Analyzing Multi-agent Task Plans for Intelligent Virtual Training System using Petri Nets

Integrated virtual reality with intelligent tutoring system, a multi-agent architecture was proposed for intelligent virtual training system (IVTS) for mine safety training. In order to make sure IVTS agents task plans reliable and adaptive, a Petri nets-based declarative method was applied to model the virtual training task planning knowledge, which was represented as task planning knowledge Petri nets (TP-PNets), and an algorithm was implemented to construct TP-PNets. Then, hierarchy colored Petri nets (HCPN) was used to model multi-agent task planning behaviors for IVTS, and simulation and message sequence chart was used to analyze and verify the agent task planning HCPN model

Virtual Sensor Modeling for Multi-agent Intelligent Virtual Environment

Every agent in the proposed intelligent virtual environment collects information through sensors to percept on environment and to interact with others, which puts emphasis on agent sensing and sensory data representation. The paper focuses on developing agent virtual sensors model to represent agent sensory information. A virtual distance sensor model was realized to navigate unmanned vehicle in virtual mine training environment by analysing the depth buffer information of the rendered scene and a laser range finder. Integrated the virtual entity physical attributes, the virtual environment atmospheric model and the real sensor properties, the virtual sensor model can simulate the high-fidelity sensor functions to provide trainee with sufficient visual information for decision-making training.

Multi-agent based intelligent virtual training system

Integrated virtual reality with intelligent tutoring system, a multi-agent framework was proposed for intelligent virtual system for training in optional environment to improve traineepsilas emergency decision-making ability as well as their technical level. In order to make sure agentpsilas training task plans are reliable and adaptive, a declarative formalism was applied to represent the virtual training task knowledge. Moreover, to mediate the interactions between trainee and virtual environment, the architecture of virtual trainee agent was constructed and realized, which is a comprehensive model of 3D virtual human as highly-capable individuals and combines with perceptual, behavioral, mental state, and cognitive modules. Finally, a practical intelligent virtual training system was represented, and the primary results show the robustness of our system and its ability to produce an effective virtual environment for safety training.

Bayesian Decision-Theoretic Causation Analysis for Gas Explosion

Frequent mine gas incident in China fully shows there must be many undiscovered essential factors in all previous gas incident investigations. Exploring a new analysis approach is vital to prevent and control mine gas emergency. A Bayesian decision-theoretic causation analysis method was introduced to study the mine gas incident. A typical gas explosion case was explained. The interaction law of human, machine, environment, and management factors causing mine gas incident was discussed. The evolution process of the connotative factors from the uncertain state to the certain state was discovered, and the relationships between the external factors and the internal factors were revealed