基于ECA规则实现PVR系统中任务演绎和动作规划的新方法

Implementation for task deduction and action planning in PVR system using ECA rules

投射式虚拟现实(PVR)技术是一种用来高效控制和管理机器人及其他自动化设备的新方法,任务演绎和动作规划是其中两大关键部件。从目前应用情况看,这两大部件在非结构化和部分未知环境下的执行过程中还存在着难以被有效构造等不足。通过将主动数据库中的ECA规则引入到基于PVR技术构建的仿真系统,讨论了实现任务演绎和动作规划的新方法,并据此建立了一个新的软件框架TD&AP_ECA。该框架不仅具有良好的柔韧性(即能根据操作者的要求方便地重新构造部分参数化仿真系统),而且为PVR技术的协作式和分布式应用提供了良好的扩展接口。通过机械手抓物的真实机器人仿真实例,形象演示了基于ECA规则的任务演绎和动作规划部件的柔性处理过程。同时,借助基于CORBA的EJB技术为TD&AP_ECA的分布式应用描述和实现了相应的接口。

Projective virtual reality （PVR） is a new approach used to control and supervise robots and other automation equipments. Task deduction （TD） and action planning （AP） are two of the key components of PVR technology. In this paper, in order to overcome the deficiency of TD and AP which is difficult to be constructed in unstructured and less known environments, the event- condition - action （ECA） rules of active database （ADB） are introduced into PVR - based simulation system to implement the TD and AP. Then, a new software framework named TD＆AP_ ECA is constructed. Framework TD＆AP_ ECA not only has favorable flexibility （ namely, it can re - configure the PVR system according to operators＇ requirements）, but also provides an excellent application interface for collaborative and distributed applications of PVR. To demonstrate how to realize the flexible treatment to TD and AP by means of ECA rules, a case study involving in grasping ball by a manipulator is presented in this paper. Meanwhile, the well - behaved CORBA- based enterprise JavaBeansTM（EJB） technology is used to describe the interface for distributed implementation of TD＆AP_ ECA.