WO2017008752A1 - 一种机器人指令处理方法、装置及机器人 - Google Patents
一种机器人指令处理方法、装置及机器人 Download PDFInfo
- Publication number
- WO2017008752A1 WO2017008752A1 PCT/CN2016/090023 CN2016090023W WO2017008752A1 WO 2017008752 A1 WO2017008752 A1 WO 2017008752A1 CN 2016090023 W CN2016090023 W CN 2016090023W WO 2017008752 A1 WO2017008752 A1 WO 2017008752A1
- Authority
- WO
- WIPO (PCT)
- Prior art keywords
- instruction
- local
- conditional reflection
- reflection
- conditional
- Prior art date
Links
Images
Classifications
-
- G—PHYSICS
- G05—CONTROLLING; REGULATING
- G05B—CONTROL OR REGULATING SYSTEMS IN GENERAL; FUNCTIONAL ELEMENTS OF SUCH SYSTEMS; MONITORING OR TESTING ARRANGEMENTS FOR SUCH SYSTEMS OR ELEMENTS
- G05B19/00—Programme-control systems
- G05B19/02—Programme-control systems electric
- G05B19/04—Programme control other than numerical control, i.e. in sequence controllers or logic controllers
Definitions
- the present invention relates to the field of robot technology, and in particular, to a robot instruction processing method, apparatus, and robot.
- the realization of the robot basically adopts a solution integrating the sensing system, the cognitive system and the control system, wherein the cognitive system is the core system of the robot, which is equivalent to the human brain.
- the robot's sensing system, cognitive system and control system are placed on the robot to realize the human-like work.
- the cognitive system that is, the brain of the robot
- the cognitive system is placed locally on the robot. Due to the development of current computer and artificial intelligence technologies, and the local imitating human brain requires a huge amount of computing resources, it is subject to local intelligence.
- the embodiment of the invention provides a robot instruction processing method, device and a robot to solve the technical problem that the robot scheme in the prior art is limited to local resources and can only implement simple tasks.
- an embodiment of the present invention provides a robot instruction processing method, including the following steps:
- the type of the instruction including an unconditional instruction and a conditional inverse Shooting instruction
- If the instruction is a conditional reflection instruction, querying the conditional reflection instruction in a local instruction set;
- conditional reflection instruction If the conditional reflection instruction does not exist in the local instruction set, the conditional reflection instruction is sent to the cloud.
- an embodiment of the present invention provides a robot instruction processing apparatus, including:
- a receiving module configured to receive a user instruction
- a determining module configured to determine a type of the instruction, the type of the instruction includes a non-conditional reflection instruction and a conditional reflection instruction;
- a query module configured to query the conditional reflection instruction in a local instruction set if the instruction is a conditional reflection instruction
- a sending module configured to send the conditional reflection instruction to the cloud if the conditional reflection instruction does not exist in the local instruction set.
- an embodiment of the present invention provides a robot comprising: the robot instruction processing device as described above.
- an embodiment of the present invention provides a robot, including: a processor and a storage medium; and the storage medium stores instructions for executing the foregoing method.
- the robot instruction processing method, apparatus and robot provided by the embodiments of the present invention first determine the type of the user instruction after receiving the user instruction, and the instruction type may include a non-conditional reflection instruction and a conditional reflection instruction, and the condition reflection instruction may be used for the conditional reflection instruction.
- the query is first performed in the local instruction set. If the conditional reflection instruction does not exist in the local instruction set, the conditional reflection instruction is sent to the cloud.
- the embodiment of the present invention after receiving the user instruction, determining the type of the instruction, querying the local instruction set for the conditional reflection instruction, and transmitting the conditional reflection instruction that does not exist in the local instruction set to the cloud, and processing by the cloud , so you don't have local resources
- the limitations allow the robot to perform more complex tasks, further enhancing the robot's human-like ability.
- FIG. 1 is a schematic flow chart showing an implementation of a robot instruction processing method in an embodiment of the present invention
- FIG. 2 is a schematic flowchart diagram of a method for generating a local instruction set in an embodiment of the present invention
- FIG. 3 is a schematic structural diagram of a robot instruction processing apparatus in an embodiment of the present invention.
- some robots are implemented by putting a cognitive system on a cloud platform, and the cloud platform is basically used for robot management, coordination, and the like, and does not undertake other tasks.
- the inventor believes that for the above-mentioned robots, if all the task instructions are handed over to the cloud platform for processing, not only the cloud platform resources will be wasted, but also the communication resources of the system will be wasted.
- the embodiment of the present invention provides a robot instruction processing method and apparatus, which will be described below.
- FIG. 1 is a schematic flowchart showing the implementation of a robot instruction processing method in an embodiment of the present invention. As shown in the figure, the robot instruction processing method may include the following steps:
- Step 101 Receive a user instruction.
- Step 102 Determine a type of the user instruction, where the type of the user instruction includes a non-conditional reflection instruction and a conditional reflection instruction;
- Step 103 If the user instruction is a conditional reflection instruction, querying the conditional reflection instruction in a local instruction set;
- Step 104 If the conditional reflection instruction does not exist in the local instruction set, send the conditional reflection instruction to the cloud.
- the user instruction may be received by the sensing system of the robot, and the user instruction may include an instruction received by means of voice, remote control, gesture, touch, button, and the like.
- these instructions can be classified in advance, and the instructions can be divided into non-conditional reflection instructions and conditional reflection instructions according to the human system.
- the non-conditioned reflection instruction can also be called an unconditional reflection instruction, which can be understood as an instruction that can be processed without the need for robot conditional reflection. For example, after the user presses a certain button, the button itself has a certain meaning, and the robot does not need to perform the condition. The reflection can be processed accordingly and the corresponding operation can be performed.
- non-conditioned reflection instructions can be understood as instructions that can cause reflective reactions, innate reflection commands, and instinct reflections without the need for acquired training.
- conditional reflection instruction that is, the instruction that the robot needs to be conditionally reflected
- the conditional reflection instruction can be understood as the reflection instruction gradually formed after the day after learning, for example, the user sends a piece of speech, and the robot receives this.
- segment voice command it is necessary to first perform speech recognition and analysis on the speech (which can be understood as a conditional reflection process). After such a reflection time, the meaning of the instruction can be understood finally, and the corresponding instruction processing and execution are performed. Corresponding operations, etc.
- the type of the instruction is judged, and then the instructions may be separately processed according to the type of the instruction.
- the instructions may be separately processed according to the type of the instruction.
- non-conditioned reflection instructions since direct operation is not required for conditional reflection, it can be directly processed locally, similar to human processing of unconditioned reflections.
- the conditional reflection instruction it is first necessary to search in the local instruction set to determine whether the instruction is in the local instruction set. If there is, it can be directly processed locally, if not Yes, it needs to be sent to the cloud and processed by the cloud.
- the embodiment of the present invention does not process all the commands sent by the user locally in the robot, nor does it send all the user instructions to the cloud for processing, but first classifies the user instructions, and part of the local direct processing and another part is sent to Cloud processing makes the instruction processing of the robot no longer limited to local resources, so that the scope of tasks that the robot can achieve can be expanded, and more complicated tasks can be completed, and at the same time, since the embodiment of the present invention only does not exist the local instruction set.
- the conditional reflection instruction is sent to the cloud, which greatly reduces the waste of communication resources.
- the conditional reflection instructions existing in the local instruction set are still directly processed by the local, reducing the waste of cloud resources.
- the non-conditioned reflection instruction may include a remote control instruction, a touch instruction, and/or a key instruction; the conditional reflection instruction may include a voice instruction, a gesture instruction, an eye instruction, and/or an expression instruction.
- the user instruction divides the user instruction, and classifies the instructions used by the traditional home appliance class, such as: a remote control command, a touch command, a button command, etc., into an unconditioned instruction; the voice, the gesture, the eyes, Instructions received by input methods such as expressions are classified as conditional reflection instructions.
- this classification method is based on the human system, the classification method can achieve a better humanoid effect.
- the data structure of the local instruction set may include an instruction execution time, an instruction usage interval, and an instruction content.
- the data result of the local instruction set may be an array, a queue, a stack, a graph, and the like.
- the content of each element in the data structure may include: an instruction execution time, an instruction usage interval, and an instruction content.
- the instruction usage interval can be calculated by using the instruction timestamp.
- the method may further include:
- conditional reflection instruction exists in the local instruction set, the conditional reflection instruction is locally processed, and the instruction execution time of the conditional reflection instruction is incremented by one, and the instruction use interval of the conditional reflection instruction is updated.
- conditional reflection instruction when the conditional reflection instruction exists in the local instruction set, the conditional reflection instruction may be locally processed, and the local instruction set may be updated.
- the specific update content may be: adding one instruction execution time of the conditional reflection instruction And updating the instruction use interval of the conditional reflection instruction.
- conditional reflection instruction existing in the local instruction set can be understood as the same as the non-conditioned reflection instruction, which has the advantages of faster processing speed, higher efficiency, and no need to send to the cloud, thereby saving communication resources and the like.
- the method may further include:
- the local instruction set has a space, storing the instruction content of the conditional reflection instruction to the local instruction set, initializing the instruction execution number of the conditional reflection instruction to 1, and adding an instruction to the conditional reflection instruction. Timestamp
- the conditional reflection instruction can be sent to the cloud for processing by the cloud, and after receiving the instruction processing result in the cloud, the robot can control the robot to perform corresponding operations, and can check the local Whether the instruction set has space, if any, store the instruction content to the local instruction set, and initialize the instruction execution count to 1, increase the instruction use timestamp; if the instruction set space is full, the local instruction set has the existing instruction Concentrate to find the instructions that use the least number of times or use the longest interval, with new instructions (the strips Replace the instruction with a reflection command).
- the instruction with the least number of uses or the longest use interval may be set according to actual needs, or may be set in advance by setting a related threshold, or setting a priority, etc., which is not limited by the present invention.
- the embodiment of the invention provides an algorithm based on self-learning update iteration, which can make the local instruction set update itself every time the instruction is executed, and provide efficiency guarantee for subsequent instruction processing.
- the local sensing system of the robot receives user commands, which may include: voice commands, remote commands, gestures, touches, buttons, and the like.
- the instructions can be divided into non-conditioned reflection instructions and conditional reflection instructions.
- the classification method can be various, for example, the instructions (the remote control command, the touch command, the key command, etc.) used in the conventional home appliance class are classified as non-conditional reflection instructions; the instructions received in the input manners such as voice, gesture, eyes, and expressions are received. Classified as a conditional reflection instruction.
- non-conditioned reflection instructions For non-conditioned reflection instructions, it is directly processed locally, and the specific processing is similar to human processing of non-conditioned reflection instructions;
- conditional reflection instruction it is first determined whether the instruction is in the local instruction set, and if so, directly processed locally; if not, it is sent to the cloud for processing.
- conditional reflection instruction existing in the local instruction set can be understood as having the same advantages as the non-conditional reflection instruction, that is, the processing speed is faster, the efficiency is higher, and the communication resource is saved without being sent to the cloud.
- the local conditional reflection instruction set may be based on a self-learning update iteration mode.
- FIG. 2 is a schematic flowchart diagram of a local instruction set generation method in the embodiment of the present invention, which will be described below.
- Step 201 Establish a data structure of a local instruction set.
- the content of each element in the data structure may include: the number of times the instruction is received, the average interval of instruction usage, and the content of the instruction.
- Step 202 Initialize a local instruction set.
- the initialization of the local instruction set may be empty, or some common instructions may be preset, for example, a voice wake-up command, a voice shutdown command, a gesture shutdown command, etc., and the number of command receptions and the average use interval of the commands may also be preset.
- step 203 If yes, go to step 203;
- step 204 is performed.
- Step 203 Query a local instruction set.
- the robot After receiving the conditional reflection instruction, the robot first queries whether the instruction is in the local instruction set:
- step 204 is performed;
- step 205 If not, proceed to step 205;
- Step 204 Local direct processing, execution instruction execution time plus 1, update instruction usage time interval;
- Step 205 Send the instruction to the cloud (the cognitive system of the robot) for processing.
- the robot's control system can control the robot to perform corresponding operations, and check whether there is space in the local instruction set:
- step 206 is performed;
- step 207 is performed.
- Step 206 Store the instruction directly.
- the instruction is stored in the local instruction set, and step 208 is performed;
- Step 207 Replace the existing instruction with the instruction.
- Step 208 Initialize the instruction execution number of the instruction to 1, and increase the instruction use time stamp.
- a robot instruction processing device is also provided in the embodiment of the present invention. Since the principle of solving the problem of these devices is similar to a robot instruction processing method, the implementation of these devices can be referred to the implementation of the method, and the repetition is performed. No longer.
- FIG. 3 is a schematic structural diagram of a robot instruction processing apparatus according to an embodiment of the present invention. As shown in the figure, the robot instruction processing apparatus may include:
- the receiving module 301 is configured to receive a user instruction
- the determining module 302 is configured to determine a type of the instruction, and the type of the instruction includes a non-conditional reflection instruction and a conditional reflection instruction;
- the query module 303 is configured to: if the instruction is a conditional reflection instruction, query the conditional reflection instruction in a local instruction set;
- the sending module 304 is configured to send the conditional reflection instruction to the cloud if the conditional reflection instruction does not exist in the local instruction set.
- the non-conditioned reflection instruction may include a remote control instruction, a touch instruction, and/or a key instruction; the conditional reflection instruction may include a voice instruction, a gesture instruction, an eye instruction, and/or an expression instruction.
- the data structure of the local instruction set may include an instruction execution time, an instruction usage interval, and an instruction content.
- the apparatus may further include:
- a local processing module configured to locally process the conditional reflection instruction if the conditional reflection instruction exists in the local instruction set
- an update module configured to increase an instruction execution time of the conditional reflection instruction by one, and update an instruction usage interval of the conditional reflection instruction.
- a space checking module configured to check whether there is space in the local instruction set
- An instruction storage module configured to: if the local instruction set has a space, store the instruction content of the conditional reflection instruction to the local instruction set, and initialize the instruction execution number of the conditional reflection instruction to 1, The conditional reflection instruction increases the instruction using a timestamp;
- An instruction replacement module configured to: if the local instruction set space is full, find an instruction that has the least number of executions of the instruction in the local instruction set or the instruction usage interval, and replaces the conditional reflection instruction with the instruction execution number minimum or The instruction uses the instruction with the longest interval.
- an embodiment of the present invention further provides a robot, including: the robot instruction processing device as described above.
- an embodiment of the present invention provides a robot, including: a processor and a storage medium; and the storage medium stores instructions for executing the foregoing method.
- embodiments of the present invention can be provided as a method, system, or computer program product. Accordingly, the present invention may take the form of an entirely hardware embodiment, an entirely software embodiment, or a combination of software and hardware. Moreover, the invention can take the form of a computer program product embodied on one or more computer-usable storage media (including but not limited to disk storage, CD-ROM, optical storage, etc.) including computer usable program code.
- computer-usable storage media including but not limited to disk storage, CD-ROM, optical storage, etc.
- the computer program instructions can also be stored in a computer readable memory that can direct a computer or other programmable data processing device to operate in a particular manner, such that the instructions stored in the computer readable memory produce an article of manufacture comprising the instruction device.
- the apparatus implements the functions specified in one or more blocks of a flow or a flow and/or block diagram of the flowchart.
- These computer program instructions can also be loaded onto a computer or other programmable data processing device such that a series of operational steps are performed on a computer or other programmable device to produce computer-implemented processing for execution on a computer or other programmable device.
- the instructions provide steps for implementing the functions specified in one or more of the flow or in a block or blocks of a flow diagram.
Abstract
提供了一种机器人指令处理方法及装置，包括：接收用户指令（101）；判断该指令的类型，该指令的类型包括非条件反射指令和条件反射指令（102）；如果该指令为条件反射指令，将该条件反射指令在本地指令集中进行查询（103）；若该本地指令集中不存在该条件反射指令，则将该条件反射指令发送至云端（104）。该方法接收到用户指令并判断所述用户指令的类型后，将本地指令集中不存在的条件反射指令则发送至云端，由云端处理，这样即可不受本地资源的限制，从而使得机器人可以完成更加复杂的任务，进一步提高了机器人的仿人能力。
Description
本发明涉及机器人技术领域，特别涉及一种机器人指令处理方法、装置及机器人。
现有技术中，机器人的实现基本是采用了感知系统、认知系统、控制系统一体化的解决方案，其中，认知系统是机器人的核心系统，相当于人的大脑。在具体实现时，将机器人的感知系统、认知系统和控制系统均放在机器人本地，来实现仿人工作。
上述方式中，将认知系统(也就是，机器人的大脑)放在机器人本地，由于受到当前计算机以及人工智能技术的发展限制，且在本地模仿人的大脑需要巨量的运算资源，受到本地智能仿人机器的体积、功耗、移动性等方面的限制，导致现有机器人只能实现一些简单的任务，无法完成复杂的任务。
现有技术不足在于：
现有机器人方案受限于本地资源，只能实现简单任务。
发明内容
本发明实施例提出了一种机器人指令处理方法、装置及机器人，以解决现有技术中机器人方案受限于本地资源，只能实现简单任务的技术问题。
在一个方面，本发明实施例提供了一种机器人指令处理方法，包括如下步骤：
接收用户指令；
判断所述指令的类型，所述指令的类型包括非条件反射指令和条件反
射指令；
如果所述指令为条件反射指令，将所述条件反射指令在本地指令集中进行查询；
若所述本地指令集中不存在所述条件反射指令，则将所述条件反射指令发送至云端。
在另一个方面，本发明实施例提供了一种机器人指令处理装置，包括：
接收模块，用于接收用户指令；
判断模块，用于判断所述指令的类型，所述指令的类型包括非条件反射指令和条件反射指令；
查询模块，用于如果所述指令为条件反射指令，将所述条件反射指令在本地指令集中进行查询；
发送模块，用于若所述本地指令集中不存在所述条件反射指令，则将所述条件反射指令发送至云端。
在另一个方面，本发明实施例提供了一种机器人，包括：如上述所述的机器人指令处理装置。
在另一个方面，本发明实施例提供了一种机器人，包括：处理器和存储介质；所述存储介质中存储有用于执行上述方法的指令。
有益效果如下：
本发明实施例所提供的机器人指令处理方法、装置及机器人，在接收到用户指令后，首先判断所述用户指令的类型，指令类型可以包括非条件反射指令和条件反射指令，对于条件反射指令可以先在本地指令集中进行查询，如果本地指令集中不存在该条件反射指令，则将该条件反射指令发送至云端。由于本发明实施例所提供的方案，在接收到用户指令后判断指令的类型，对于条件反射指令在本地指令集进行查询，将本地指令集中不存在的条件反射指令则发送至云端，由云端处理，这样即可不受本地资源
的限制，从而使得机器人可以完成更加复杂的任务，进一步提高了机器人的仿人能力。
下面将参照附图描述本发明的具体实施例，其中：
图1示出了本发明实施例中机器人指令处理方法实施的流程示意图；
图2示出了本发明实施例中本地指令集的生成方法的流程示意图；
图3示出了本发明实施例中机器人指令处理装置的结构示意图。
为了使本发明的技术方案及优点更加清楚明白，以下结合附图对本发明的示例性实施例进行进一步详细的说明，显然，所描述的实施例仅是本发明的一部分实施例，而不是所有实施例的穷举。并且在不冲突的情况下，本说明中的实施例及实施例中的特征可以互相结合。
发明人在发明过程中注意到：
在现有技术中，有一些机器人的实现方案是，将认知系统放到云平台，云平台基本上都是用于机器人管理、协同等工作，不承担其他任务。
发明人认为，对于上述机器人而言，如果将全部任务指令都交给云平台处理的话，不仅会浪费云平台资源，还会浪费系统的通信资源。
针对上述不足，本发明实施例提出了一种机器人指令处理方法及装置，下面进行说明。
图1示出了本发明实施例中机器人指令处理方法实施的流程示意图，如图所示，所述机器人指令处理方法可以包括如下步骤：
步骤101、接收用户指令；
步骤102、判断所述用户指令的类型，所述用户指令的类型包括非条件反射指令和条件反射指令；
步骤103、如果所述用户指令为条件反射指令，将所述条件反射指令在本地指令集中进行查询；
步骤104、若所述本地指令集中不存在所述条件反射指令，则将所述条件反射指令发送至云端。
在具体实施中，可以由机器人的感知系统接收用户指令，用户指令可以包括：语音、遥控、手势、触摸、按键等方式接收的指令。
在本发明实施例中可以预先将这些指令进行分类，可以仿照人的系统，将指令分为非条件反射指令和条件反射指令。
其中，非条件反射指令也可以称为无条件反射指令，可以理解为不需要机器人条件反射就可以处理的指令，例如：用户按下某个按键之后，该按键本身就具备一定含义，机器人无需进行条件反射就可以进行相应的指令处理、执行相应的操作。与人的非条件反射相对应理解，可以将非条件反射指令理解为不需要后天的训练就能引起反射性反应的指令、先天性反射指令、本能反射的指令。
相应的，条件反射指令，也即机器人需要条件反射才能处理的指令，对应人的系统的话，则可以理解为经过后天学习逐渐形成的反射指令，例如：用户发送了一段语音，机器人在接收到这段语音指令后，需要首先对这段语音进行语音识别、分析等操作(可以对应理解为条件反射过程)，在这样一段反射时间后，最终才能理解该指令的含义，进行相应的指令处理、执行相应的操作等。
在接收到用户指令后，判断所述指令的类型，然后可以根据指令的类型的不同，将指令分别进行不同的处理。对于非条件反射指令，由于不需要条件反射即可直接操作，则可以直接由本地处理，和人类处理非条件反射的指令类似。而对于条件反射指令，则需要首先在本地指令集中查找，确定本地指令集中是否有该指令，如果有则可以由本地直接处理，如果没
有，则需要发送至云端，由云端处理。
由于本发明实施例并不是将用户发送过来的所有指令均在机器人本地处理，也不是将所有用户指令均发送给云端处理，而是先将用户指令进行分类，一部分本地直接处理、另一部分发往云端处理，使得机器人的指令处理不再受限于本地资源，从而可以扩大机器人可以实现的任务范围，确保更复杂的任务也可以完成，同时由于本发明实施例仅是将本地指令集不存在的条件反射指令发送给云端，从而大大的降低了通信资源的浪费，对于本地指令集中存在的条件反射指令依然由本地直接处理，减少了云端资源的浪费。
实施中，所述非条件反射指令可以包括遥控指令、触摸指令和/或按键指令；所述条件反射指令可以包括语音指令、手势指令、眼神指令和/或表情指令。
本发明实施例根据人的系统，将用户指令进行了划分，将传统家电类使用的指令，例如：遥控指令、触摸指令、按键指令等归类为非条件反射指令；将语音、手势、眼神、表情等输入方式接收的指令归类为条件反射指令。
由于这种分类方式是按照人的系统来划分指令的，因此，采用这种分类方式，可以达到更好的仿人效果。
实施中，所述本地指令集的数据结构中可以包括指令执行次数、指令使用间隔和指令内容。
在具体实施中，本地指令集的数据结果可以为数组、队列、栈、图等等，数据结构中每个元素的内容可以包括：指令执行次数、指令使用间隔和指令内容。其中，指令使用间隔可以通过指令使用时间戳计算得到。
实施中，在所述将所述条件反射指令在本地指令集中查询之后，所述方法可以进一步包括：
若所述本地指令集中存在所述条件反射指令，则由本地处理所述条件反射指令，并将所述条件反射指令的指令执行次数加一、更新所述条件反射指令的指令使用间隔。
也即，在本地指令集存在该条件反射指令时，可以由本地处理该条件反射指令，同时可以更新所述本地指令集，具体更新内容可以为：将所述条件反射指令的指令执行次数加一、更新所述条件反射指令的指令使用间隔等。
本地指令集中存在的条件反射指令，可以理解为与非条件反射指令一样，具有处理速度更快、效率更高，而且无需发送至云端，节省了通信资源等优点。
实施中，在所述若所述本地指令集中不存在所述条件反射指令，则将所述条件反射指令发送至云端之后，所述方法可以进一步包括：
在收到云端的指令处理结果后，检查所述本地指令集是否还有空间；
如果所述本地指令集有空间，则将所述条件反射指令的指令内容存储到所述本地指令集，将所述条件反射指令的指令执行次数初始化为1，为所述条件反射指令增加指令使用时间戳；
如果所述本地指令集空间已满，则查找所述本地指令集中指令执行次数最少或指令使用间隔最长的指令，将所述条件反射指令替换所述指令执行次数最少或指令使用间隔最长的指令。
也即，如果本地指令集没有该条件反射指令，则可以将该条件反射指令发送至云端，由云端进行处理，机器人接收到云端的指令处理结果后，可以控制机器人进行相应操作，同时可以检查本地指令集是否有空间，如果有，则将指令内容存储到本地指令集，同时将指令执行次数初始化为1、增加指令使用时间戳；如果指令集空间已满，则将本地指令集已有的指令集中，找出使用次数最少或使用时间间隔最长的指令，用新指令(所述条
件反射指令)替换此指令。
在具体实施中，使用次数最少或使用时间间隔最长的指令，可以根据实际需要进行设定，也可以预先设置相关的阈值，或者设定优先级等方式实现，本发明对此不作限制。
本发明实施例提供了一种基于自学习更新迭代的算法，可以使得本地指令集在每次指令执行时均可以自行更新，为后续的指令处理提供效率保障。
为了便于本发明的实施，下面以实例进行说明。
机器人本地的感知系统收到用户指令，这些指令可以包括：语音指令、遥控指令、手势、触摸、按键等等。
将这些指令进行分类，可以将指令分为非条件反射指令和条件反射指令。分类方法可以多种多样，例如：将传统家电类使用的指令(遥控指令、触摸指令、按键指令等)归类为非条件反射指令；将语音、手势、眼神、表情等输入方式接收到的指令归类为条件反射指令。
将这些指令进行指令预处理，具体预处理过程可以如下：
判断指令的类型，根据类型的不同，将指令发送至不同的处理模块：
对于非条件反射指令，直接由本地处理，具体处理过程与人类处理非条件反射指令类似；
对于条件反射指令，首先判断本地指令集中是否有该指令，如果有，由本地直接处理；如果没有，则发送至云端处理。
本地指令集中存在的条件反射指令，可以理解为与非条件反射指令具有一样的优点，也即，处理速度更快、效率更高、且无需发送至云端节省了通信资源。在具体实施中，本地条件反射指令集可以基于自学习更新迭代方式，图2示出了本发明实施例中本地指令集的生成方法的流程示意图，下面进行说明。
步骤201、建立本地指令集的数据结构。
建立一个本地指令集的数据结构，其数据结构可以是数组、队列、栈、图等等。其中，数据结构中每个元素的内容可以包括：指令接收次数、指令平均使用间隔和指令内容等。
步骤202、初始化本地指令集。
本地指令集的初始化可以为空，也可以预置一些常用指令，例如：语音唤醒指令、语音关机指令、手势关机指令等等，其指令接收次数、指令平均使用间隔也都可以预先设置。
然后，判断是否为条件反射指令：
如果是，则执行步骤203；
如果不是，则执行步骤204。
步骤203、查询本地指令集。
机器人在收到条件反射指令后，首先查询本地指令集中是否有该指令：
如果存在该指令，则执行步骤204；
如果没有，则执行步骤205；
步骤204、本地直接处理，执行指令执行次数加1、更新指令使用时间间隔；
步骤205、将该指令发送至云端(机器人的认知系统)处理。
在收到认知系统的指令处理结果后，机器人的控制系统可以控制机器人进行相应的操作，同时检查本地指令集是否还有空间：
如果有空间，则执行步骤206；
如果没有空间，则执行步骤207。
步骤206、直接存储该指令。
将该指令存储到本地指令集，执行步骤208；
步骤207、用该指令替换已有指令。
如果本地指令集空间已满，可以将本地指令集中已存在的指令集中，找出使用次数最少或使用间隔最长的指令，用新指令将其替换掉。
步骤208、将该指令的指令执行次数初始化为1、增加指令使用时间戳。
基于同一发明构思，本发明实施例中还提供了一种机器人指令处理装置，由于这些设备解决问题的原理与一种机器人指令处理方法相似，因此这些设备的实施可以参见方法的实施，重复之处不再赘述。
图3示出了本发明实施例中机器人指令处理装置的结构示意图，如图所示，机器人指令处理装置可以包括：
接收模块301，用于接收用户指令；
判断模块302，用于判断所述指令的类型，所述指令的类型包括非条件反射指令和条件反射指令；
查询模块303，用于如果所述指令为条件反射指令，将所述条件反射指令在本地指令集中进行查询；
发送模块304，用于若所述本地指令集中不存在所述条件反射指令，则将所述条件反射指令发送至云端。
实施中，所述非条件反射指令可以包括遥控指令、触摸指令和/或按键指令；所述条件反射指令可以包括语音指令、手势指令、眼神指令和/或表情指令。
实施中，所述本地指令集的数据结构中可以包括指令执行次数、指令使用间隔和指令内容。
实施中，所述装置可以进一步包括：
本地处理模块，用于若所述本地指令集中存在所述条件反射指令，本地处理所述条件反射指令；
更新模块，用于将所述条件反射指令的指令执行次数加一、更新所述条件反射指令的指令使用间隔。
实施中，所述接收模块可以进一步用于接收云端返回的指令处理结果；所述装置可以进一步包括：
空间检查模块，用于检查所述本地指令集中是否还有空间；
指令存储模块，用于如果所述本地指令集有空间，则将所述条件反射指令的指令内容存储到所述本地指令集，将所述条件反射指令的指令执行次数初始化为1，为所述条件反射指令增加指令使用时间戳；
指令替换模块，用于如果所述本地指令集空间已满，则查找所述本地指令集中指令执行次数最少或指令使用间隔最长的指令，将所述条件反射指令替换所述指令执行次数最少或指令使用间隔最长的指令。
基于同一发明构思，本发明实施例中还提供了一种机器人，包括：如上述所述的机器人指令处理装置。
基于同一发明构思，本发明实施例提供了一种机器人，包括：处理器和存储介质；所述存储介质中存储有用于执行上述方法的指令。
为了描述的方便，以上所述装置的各部分以功能分为各种模块或单元分别描述。当然，在实施本发明时可以把各模块或单元的功能在同一个或多个软件或硬件中实现。
本领域内的技术人员应明白，本发明的实施例可提供为方法、系统、或计算机程序产品。因此，本发明可采用完全硬件实施例、完全软件实施例、或结合软件和硬件方面的实施例的形式。而且，本发明可采用在一个或多个其中包含有计算机可用程序代码的计算机可用存储介质(包括但不限于磁盘存储器、CD-ROM、光学存储器等)上实施的计算机程序产品的形式。
本发明是参照根据本发明实施例的方法、设备(系统)、和计算机程序产品的流程图和/或方框图来描述的。应理解可由计算机程序指令实现流程图和/或方框图中的每一流程和/或方框、以及流程图和/或方框图中
的流程和/或方框的结合。可提供这些计算机程序指令到通用计算机、专用计算机、嵌入式处理机或其他可编程数据处理设备的处理器以产生一个机器，使得通过计算机或其他可编程数据处理设备的处理器执行的指令产生用于实现在流程图一个流程或多个流程和/或方框图一个方框或多个方框中指定的功能的装置。
这些计算机程序指令也可存储在能引导计算机或其他可编程数据处理设备以特定方式工作的计算机可读存储器中，使得存储在该计算机可读存储器中的指令产生包括指令装置的制造品，该指令装置实现在流程图一个流程或多个流程和/或方框图一个方框或多个方框中指定的功能。
这些计算机程序指令也可装载到计算机或其他可编程数据处理设备上，使得在计算机或其他可编程设备上执行一系列操作步骤以产生计算机实现的处理，从而在计算机或其他可编程设备上执行的指令提供用于实现在流程图一个流程或多个流程和/或方框图一个方框或多个方框中指定的功能的步骤。
尽管已描述了本发明的优选实施例，但本领域内的技术人员一旦得知了基本创造性概念，则可对这些实施例作出另外的变更和修改。所以，所附权利要求意欲解释为包括优选实施例以及落入本发明范围的所有变更和修改。
Claims (12)
- 一种机器人指令处理方法，其特征在于，包括如下步骤：接收用户指令；判断所述用户指令的类型，所述用户指令的类型包括非条件反射指令和条件反射指令；如果所述用户指令为条件反射指令，将所述条件反射指令在本地指令集中进行查询；若所述本地指令集中不存在所述条件反射指令，则将所述条件反射指令发送至云端。
- 如权利要求1所述的方法，其特征在于，所述非条件反射指令包括遥控指令、触摸指令和/或按键指令；所述条件反射指令包括语音指令、手势指令、眼神指令和/或表情指令。
- 如权利要求1所述的方法，其特征在于，所述本地指令集的数据结构中包括指令执行次数、指令使用间隔和指令内容。
- 如权利要求3所述的方法，其特征在于，在所述将所述条件反射指令在本地指令集中查询之后，进一步包括：若所述本地指令集中存在所述条件反射指令，则由本地处理所述条件反射指令，并将所述条件反射指令的指令执行次数加一、更新所述条件反射指令的指令使用间隔。
- 如权利要求3所述的方法，其特征在于，在所述若所述本地指令集中不存在所述条件反射指令，则将所述条件反射指令发送至云端之后，进一步包括：在收到云端的指令处理结果后，检查所述本地指令集是否还有空间；如果所述本地指令集有空间，则将所述条件反射指令的指令内容存储到所述本地指令集，将所述条件反射指令的指令执行次数初始化为1，为所 述条件反射指令增加指令使用时间戳；如果所述本地指令集空间已满，则查找所述本地指令集中指令执行次数最少或指令使用间隔最长的指令，将所述条件反射指令替换所述指令执行次数最少或指令使用间隔最长的指令。
- 一种机器人指令处理装置，其特征在于，包括：接收模块，用于接收用户指令；判断模块，用于判断所述指令的类型，所述指令的类型包括非条件反射指令和条件反射指令；查询模块，用于如果所述指令为条件反射指令，将所述条件反射指令在本地指令集中进行查询；发送模块，用于若所述本地指令集中不存在所述条件反射指令，则将所述条件反射指令发送至云端。
- 如权利要求6所述的装置，其特征在于，所述非条件反射指令包括遥控指令、触摸指令和/或按键指令；所述条件反射指令包括语音指令、手势指令、眼神指令和/或表情指令。
- 如权利要求6所述的装置，其特征在于，所述本地指令集的数据结构中包括指令执行次数、指令使用间隔和指令内容。
- 如权利要求8所述的装置，其特征在于，进一步包括：本地处理模块，用于本地指令集中存在所述条件反射指令时，本地处理所述条件反射指令；更新模块，用于将所述条件反射指令的指令执行次数加一、更新所述条件反射指令的指令使用间隔。
- 如权利要求8所述的装置，其特征在于，所述接收模块进一步用于接收云端返回的指令处理结果；所述装置进一步包括：空间检查模块，用于检查所述本地指令集中是否还有空间；指令存储模块，用于如果所述本地指令集有空间，则将所述条件反射指令的指令内容存储到所述本地指令集，将所述条件反射指令的指令执行次数初始化为1，为所述条件反射指令增加指令使用时间戳；指令替换模块，用于如果所述本地指令集空间已满，则查找所述本地指令集中指令执行次数最少或指令使用间隔最长的指令，将所述条件反射指令替换所述指令执行次数最少或指令使用间隔最长的指令。
- 一种机器人，其特征在于，包括：如权利要求6-10任一项所述的装置。
- 一种机器人，其特征在于，包括：处理器和存储介质；所述存储介质中存储有用于执行如权利要求1-5任一项所述的方法的指令。
Applications Claiming Priority (2)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201510420192.6A CN105159148B (zh) | 2015-07-16 | 2015-07-16 | 一种机器人指令处理方法及装置 |
| CN201510420192.6 | 2015-07-16 |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| WO2017008752A1 true WO2017008752A1 (zh) | 2017-01-19 |
Family
ID=54800035
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| PCT/CN2016/090023 WO2017008752A1 (zh) | 2015-07-16 | 2016-07-14 | 一种机器人指令处理方法、装置及机器人 |
Country Status (2)
| Country | Link |
|---|---|
| CN (1) | CN105159148B (zh) |
| WO (1) | WO2017008752A1 (zh) |
Families Citing this family (10)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN105159148B (zh) * | 2015-07-16 | 2017-11-10 | 深圳前海达闼科技有限公司 | 一种机器人指令处理方法及装置 |
| CN105808399B (zh) * | 2016-03-14 | 2020-01-21 | 百度在线网络技术（北京）有限公司 | 一种远程调试的方法和装置 |
| CN107107340A (zh) * | 2016-04-15 | 2017-08-29 | 深圳前海达闼云端智能科技有限公司 | 一种云计算机器人控制装置、认知平台及控制方法 |
| CN106055105A (zh) * | 2016-06-02 | 2016-10-26 | 上海慧模智能科技有限公司 | 机器人和人机交互系统 |
| CN106142087A (zh) * | 2016-08-10 | 2016-11-23 | 东北大学 | 一种基于云计算的智能机器人系统及其控制方法 |
| CN107077578A (zh) * | 2016-12-28 | 2017-08-18 | 深圳前海达闼云端智能科技有限公司 | 机器人指令的处理方法、装置及电子设备 |
| CN106874030B (zh) * | 2016-12-30 | 2020-10-20 | 北京光年无限科技有限公司 | 机器人操作系统环境下的应用内指令解析优化方法和装置 |
| CN109070357B (zh) * | 2017-03-20 | 2022-04-15 | 深圳配天智能技术研究院有限公司 | 工业机器人系统、控制系统及方法、控制器及计算设备 |
| CN109262626A (zh) * | 2018-10-22 | 2019-01-25 | 深圳和呈睿国际技术有限公司 | 陪伴机器人控制方法、系统及移动终端 |
| CN110457119A (zh) * | 2019-07-19 | 2019-11-15 | 上海易点时空网络有限公司 | 异步处理回调方法及装置 |
Citations (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN102306234A (zh) * | 2011-06-30 | 2012-01-04 | 南京信息工程大学 | 一种气象科普机器人系统 |
| US20120316676A1 (en) * | 2011-06-10 | 2012-12-13 | Microsoft Corporation | Interactive robot initialization |
| CN103179157A (zh) * | 2011-12-22 | 2013-06-26 | 张殿礼 | 一种智能网络机器人及控制方法 |
| CN104102346A (zh) * | 2014-07-01 | 2014-10-15 | 华中科技大学 | 一种家用信息采集和用户情感识别设备及其工作方法 |
| CN104635574A (zh) * | 2014-12-15 | 2015-05-20 | 山东大学 | 一种面向幼儿的早教陪护机器人系统 |
| CN105159148A (zh) * | 2015-07-16 | 2015-12-16 | 深圳前海达闼科技有限公司 | 一种机器人指令处理方法及装置 |
Family Cites Families (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| KR20010095514A (ko) * | 2000-04-10 | 2001-11-07 | 김성헌 | 원격제어에 의한 의료 기능을 갖는 완구 시스템 |
| KR20040042242A (ko) * | 2002-11-13 | 2004-05-20 | 삼성전자주식회사 | 홈서버를 이용하는 홈로봇 및 이를 포함하는 홈네트워크시스템 |
| CN103064377A (zh) * | 2012-12-20 | 2013-04-24 | 南京航空航天大学 | 一种3g警用移动机器人双余度控制系统 |
| KR102019719B1 (ko) * | 2013-01-17 | 2019-09-09 | 삼성전자 주식회사 | 영상처리장치 및 그 제어방법, 영상처리 시스템 |
| CN204332356U (zh) * | 2014-12-30 | 2015-05-13 | 杭州微纳科技有限公司 | 一种语音命令识别装置 |
-
2015
- 2015-07-16 CN CN201510420192.6A patent/CN105159148B/zh active Active
-
2016
- 2016-07-14 WO PCT/CN2016/090023 patent/WO2017008752A1/zh active Application Filing
Patent Citations (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US20120316676A1 (en) * | 2011-06-10 | 2012-12-13 | Microsoft Corporation | Interactive robot initialization |
| CN102306234A (zh) * | 2011-06-30 | 2012-01-04 | 南京信息工程大学 | 一种气象科普机器人系统 |
| CN103179157A (zh) * | 2011-12-22 | 2013-06-26 | 张殿礼 | 一种智能网络机器人及控制方法 |
| CN104102346A (zh) * | 2014-07-01 | 2014-10-15 | 华中科技大学 | 一种家用信息采集和用户情感识别设备及其工作方法 |
| CN104635574A (zh) * | 2014-12-15 | 2015-05-20 | 山东大学 | 一种面向幼儿的早教陪护机器人系统 |
| CN105159148A (zh) * | 2015-07-16 | 2015-12-16 | 深圳前海达闼科技有限公司 | 一种机器人指令处理方法及装置 |
Also Published As
| Publication number | Publication date |
|---|---|
| CN105159148A (zh) | 2015-12-16 |
| CN105159148B (zh) | 2017-11-10 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| WO2017008752A1 (zh) | 一种机器人指令处理方法、装置及机器人 | |
| KR102379954B1 (ko) | 화상처리장치 및 방법 | |
| US11410464B2 (en) | Detection of hand gestures using gesture language discrete values | |
| WO2017128961A1 (zh) | 一种分布式系统中的训练模型的方法及设备 | |
| CA3033966A1 (en) | Utilizing a machine learning model and natural language processing to manage and allocate tasks | |
| CN108139931A (zh) | 通过重映射同步来加速任务子图 | |
| AU2018202076A1 (en) | Activity monitoring of a robot | |
| WO2018006374A1 (zh) | 一种基于主动唤醒的功能推荐方法、系统及机器人 | |
| GB2604791A (en) | Adding adversarial robustness to trained machine learning models | |
| US10608963B2 (en) | Adjusting buffer size for network interface controller | |
| CN103760970A (zh) | 一种穿戴式输入系统及输入方法 | |
| CN109955257A (zh) | 一种机器人的唤醒方法、装置、终端设备和存储介质 | |
| CN109766196A (zh) | 一种任务调度方法、装置及设备 | |
| CN110296660A (zh) | 牲畜体尺检测方法与装置 | |
| CN110488781A (zh) | 一种基于迁移强化学习的生产系统调度方法 | |
| Angleraud et al. | Coordinating shared tasks in human-robot collaboration by commands | |
| Zhong et al. | A collaborative telerobotics network framework with hand gesture interface and conflict prevention | |
| WO2017177441A1 (zh) | 一种云计算机器人控制装置、认知平台及控制方法 | |
| Ren et al. | Dynamic scheduling for energy minimization in delay-sensitive stream mining | |
| Sha et al. | Recent trends and opportunities in domain specific operating systems | |
| US20210210070A1 (en) | Skill service updating method, electronic device and readable storage medium | |
| CN109785843A (zh) | 图像处理装置和方法 | |
| CN109696910A (zh) | 一种舵机运动控制方法及装置、计算机可读存储介质 | |
| CN113821313A (zh) | 一种任务调度方法、装置及电子设备 | |
| CN111724767A (zh) | 基于狄利克雷变分自编码器的口语理解方法及相关设备 |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| 121 | Ep: the epo has been informed by wipo that ep was designated in this application |
Ref document number: 16823899 Country of ref document: EP Kind code of ref document: A1 |
|
| NENP | Non-entry into the national phase |
Ref country code: DE |
|
| 122 | Ep: pct application non-entry in european phase |
Ref document number: 16823899 Country of ref document: EP Kind code of ref document: A1 |
|
| 32PN | Ep: public notification in the ep bulletin as address of the adressee cannot be established |
Free format text: NOTING OF LOSS OF RIGHTS PURSUANT TO RULE 112(1) EPC (EPO FORM 1205A DATED 20/06/2018) |
|
| 122 | Ep: pct application non-entry in european phase |
Ref document number: 16823899 Country of ref document: EP Kind code of ref document: A1 |