发明名称 STATE AND CONTEXT DEPENDENT VOICE BASED INTERFACE FOR AN UNMANNED VEHICLE OR ROBOT
摘要 A voice-based method to control unmanned vehicles (UV) or robots that makes use of the UV or robot state and context information to constrain the output of automatic speech recognition (ASR) language models (LM) to improve the ASR accuracy and robustness in controlled and adverse environments. The voiced interaction between the human user and the machine provides a natural human-machine interface that is easy and straightforward for the human being, and reduces users' training requirements.
申请公布号 US2016161946(A1) 申请公布日期 2016.06.09
申请号 US201414501322 申请日期 2014.09.30
申请人 Speak Loud SPA 发明人 WUTH SEPULVEDA Jorge;Yoma Néstor Becerra
分类号 G05D1/00;G10L15/10;G10L15/20 主分类号 G05D1/00
代理机构 代理人
主权项 1. A voice-based method for controlling an unmanned vehicle (UV) or robot, comprising the step of: (i) preparing a finite state machine (FSM) that includes (a) a set of current possible operating states of the UV or robot,(b) a set of possible actions that the UV or robot can perform from each state in the set of current possible operating states,(c) a set of resulting possible operating states of the UV or robot resulting from each possible action of the set of possible actions performed from each state in the set of current possible operating states, and(d) one or more context variables associated with the UV or robot; (ii) providing a set of system and operating rules and logic; (iii) retrieving a present state and one or more present context variables from the UV or (iv) robot generating a list of feasible commands to which the UV or robot can respond, by employing the FSM, the present state and one or more present context variables, and the set of system and operating rules and logic; (v) acquiring a human voice command from a human user; (vi) generating an accepted voice command from the human voice command using automatic speech recognition (ASR) and a language model comprising the list of feasible commands, the accepted voice command consisting of one of the listed feasible commands; (vii) identifying an action that the UV or robot will execute based on the accepted voice command; (viii) sending a machine command to the UV or robot that effects the action that the UV will execute; and (ix) repeating steps (iii) through (ix).
地址 Santiago CL