WO2019037122A1

WO2019037122A1 - Artificial intelligence terminal and behavior control method therefor

Info

Publication number: WO2019037122A1
Application number: PCT/CN2017/099157
Authority: WO
Inventors: 孙尚传
Original assignee: 深圳市得道健康管理有限公司
Priority date: 2017-08-25
Filing date: 2017-08-25
Publication date: 2019-02-28
Also published as: CN109313450B; CN109313450A

Abstract

Disclosed is a behavior control method of an artificial intelligence terminal. The method comprises: an artificial intelligence terminal formulating at least two candidate execution policies; the artificial intelligence terminal transmitting the at least two candidate execution policies to another terminal, such that said another terminal performs evaluation of feasibilities of the respective candidate execution policies; the artificial intelligence terminal receiving the feasibilities from said another terminal; the artificial intelligence terminal selecting, according to the feasibilities, the optimal execution policy from the at least two candidate execution policies; and the artificial intelligence terminal executing the optimal execution policy. Also disclosed are an artificial intelligence terminal and a computer storage medium.

Description

Artificial intelligence terminal and behavior control method thereof

【技术领域】[Technical Field]

The present application relates to the field of computer technology, and in particular, to an artificial intelligence terminal and a behavior control method thereof.

【背景技术】【Background technique】

Artificial intelligence (AI, Artificial Intelligence) is a new technical science that studies and develops theories, methods, techniques, and applications that simulate, extend, and extend human intelligence. Artificial intelligence is a branch of computer science that attempts to understand the essence of intelligence and produce a new intelligent machine that responds in a manner similar to human intelligence. Research in this area includes robotics, speech recognition, image recognition, Natural language processing and expert systems. Since the official introduction of the artificial intelligence discipline in 1956, over the past 50 years, it has made great progress and become a wide-ranging crossover and cutting-edge science. Up to now, the development of artificial intelligence has penetrated into many aspects of social life, liberating human beings from heavy physical strength, and at the same time gradually liberating human brain labor.

The inventors of the present invention have found that the existing artificial intelligence terminals are still in the stage of weak artificial intelligence, and most of them still rely on the control of the users to perform corresponding behaviors. A few artificial intelligence terminals can perform some self-execution for external commands in combination with environmental factors. Behavior, but these behaviors are often beneficial to their own behavior, without considering the situation of other terminals, belonging to pure self-interest, contrary to human morality, and can not meet the user's requirements for artificial intelligence.

【发明内容】 [Summary of the Invention]

In order to at least partially solve the above problems, the present invention provides a method for controlling behavior of an artificial intelligence terminal, the method comprising: the artificial intelligence terminal formulating at least two candidate execution strategies; and the artificial intelligence terminal transmitting at least two candidate execution strategies to other terminals, So that other terminals respectively judge the executable degree of each candidate execution policy; the artificial intelligence terminal receives the executable degree from other terminals; the artificial intelligence terminal selects the best execution strategy from the at least two candidate execution strategies according to the executable degree Artificial intelligence terminals perform optimal execution strategies.

In order to at least partially solve the above problems, the present invention also provides an artificial intelligence terminal, the terminal comprising a processor and a communication circuit, the processor is connected to the communication circuit; the processor is configured to execute the instruction to implement the behavior of the artificial intelligence terminal as described above. Control Method.

In order to at least partially solve the above problems, the present invention also proposes a computer storage medium having a program stored therein, the program being executable to implement the behavior control method of the artificial intelligence terminal as described above.

The beneficial effects of the present invention are: the artificial intelligence terminal selects the best execution strategy and executes the best execution strategy according to the executable degree obtained by the other terminal for at least two candidate execution strategy evaluations, in addition to itself in the process of this behavior control In addition to the situation of other terminals, the behavior control based on altruistic criteria is also implemented, so that the behavioral guidelines of artificial intelligence terminals are more in line with human moral requirements, more reasonable, meet the requirements of users, and formulate implementation strategies. The process comprehensively considers the executable degree obtained by other terminal evaluations, optimizes the execution strategy of the final execution, and avoids the unnecessary overhead caused by the inappropriate execution strategy to itself and/or other terminals.

【附图说明】 [Description of the Drawings]

In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the drawings to be used in the embodiments will be briefly described below. Obviously, the drawings in the following description are only some of the present invention. For the embodiments, those skilled in the art can obtain other drawings according to the drawings without any creative work.

1 is a schematic flow chart of a first embodiment of a behavior control method of an artificial intelligence terminal according to the present invention;

2 is a schematic flow chart of a second embodiment of a behavior control method of an artificial intelligence terminal according to the present invention;

3 is a schematic flow chart of a third embodiment of a behavior control method for an artificial intelligence terminal according to the present invention;

4 is a schematic flow chart of a fourth embodiment of a behavior control method for an artificial intelligence terminal according to the present invention;

5 is a schematic flowchart of a fifth embodiment of a behavior control method of an artificial intelligence terminal according to the present invention;

6 is a schematic flowchart of a sixth embodiment of a behavior control method of an artificial intelligence terminal according to the present invention;

7 is a schematic flowchart diagram of a seventh embodiment of a behavior control method of an artificial intelligence terminal according to the present invention;

8 is a schematic flow chart of an eighth embodiment of a behavior control method for an artificial intelligence terminal according to the present invention;

9 is a schematic flow chart of a ninth embodiment of a behavior control method for an artificial intelligence terminal according to the present invention;

10 is a schematic flowchart diagram of a tenth embodiment of a behavior control method of an artificial intelligence terminal according to the present invention;

11 is a schematic flow chart of an eleventh embodiment of a behavior control method for an artificial intelligence terminal according to the present invention;

12 is a schematic flowchart diagram of a twelfth embodiment of a behavior control method for an artificial intelligence terminal according to the present invention;

13 is a schematic flowchart diagram of a thirteenth embodiment of a behavior control method of an artificial intelligence terminal according to the present invention;

14 is a schematic flowchart diagram of a fourteenth embodiment of a behavior control method of an artificial intelligence terminal according to the present invention;

15 is a schematic flowchart diagram of a fifteenth embodiment of a behavior control method of an artificial intelligence terminal according to the present invention;

16 is a schematic flowchart diagram of a sixteenth embodiment of a behavior control method of an artificial intelligence terminal according to the present invention;

17 is a schematic structural diagram of a first embodiment of an artificial intelligence terminal according to the present invention;

Figure 18 is a block diagram showing the structure of a first embodiment of the computer storage medium of the present invention.

【具体实施方式】【Detailed ways】

The technical solutions in the embodiments of the present invention will be clearly and completely described in the following with reference to the accompanying drawings. The conflicts in the following embodiments may be combined with each other. It is apparent that the described embodiments are only a part of the embodiments of the invention, and not all of the embodiments. All other embodiments obtained by a person of ordinary skill in the art based on the embodiments of the present invention without creative efforts are within the scope of the present invention.

The terms "first", "second", "third", "fourth", etc. (if present) in the specification and claims of the present invention and the above figures are used to distinguish similar objects without being used for Describe a specific order or order. It is to be understood that the data so used may be interchanged as appropriate, such that the embodiments of the invention described herein can be implemented, for example, in a sequence other than those illustrated or described herein. In addition, the terms "comprises" and "comprises" and "the" and "the" are intended to cover a non-exclusive inclusion, for example, a process, method, system, product, or device that comprises a series of steps or units is not necessarily limited to Those steps or units may include other steps or units not explicitly listed or inherent to such processes, methods, products or devices.

As shown in FIG. 1, the first embodiment of the behavior control method of the artificial intelligence terminal of the present invention includes:

S11: The artificial intelligence terminal receives an execution instruction from another terminal.

The artificial intelligence terminal may be an intelligent robot, an autonomous driving vehicle (for example, an autonomous driving vehicle) or other terminal having data analysis processing capability. Communication connection between other terminals and the terminal. Other terminals may be artificial intelligence terminals or non-intelligent terminals. The execution instructions may be given by other terminals, or may be from a user, for example, other terminals receive instructions from the user (the touch screen, keyboard, microphone, camera, mouse, etc.).

S12: The artificial intelligence terminal calculates the cost incurred by executing the execution instruction itself.

The cost overhead may refer to an increase in computation, motion, power consumption, and the like of the artificial intelligence terminal due to execution of the execution instruction and/or an additional overhead added by the user of the artificial intelligence terminal due to the execution of the execution instruction by the artificial intelligence terminal. For example, the artificial intelligence terminal is the self-driving car A, and the other terminals are the self-driving car B, both of which are driving in the same lane and A is in front of B, B wants to overtake the line and sends A to the A to temporarily move to the side lane. instruction. For A, the cost overhead is the additional cost of moving from the current lane to the side lane and back from the side lane back to the current lane.

S13: The artificial intelligence terminal determines whether the cost overhead is less than or equal to the cost benefit of other terminals.

The cost benefit is generated by other terminal terminals executing execution instructions by the artificial intelligence terminal, for example, calculations, motion, energy consumption, etc., which are reduced by other artificial terminals when executing instructions by the artificial intelligence terminal, and/or users of other terminals are executed by the artificial intelligence terminal. The benefits of the instructions. Still using the above example, for B, if A does not execute the command, then B needs to move from the current lane to the side lane over A and then from the side lane back to the current lane, otherwise if A executes the command, then B only needs to travel straight, and the difference between the two is the cost benefit of B.

The artificial intelligence terminal can calculate the cost and benefit by itself, and can also receive the cost benefit sent by other terminals.

S14: If the cost overhead is less than or equal to the cost benefit, the artificial intelligence terminal executes the execution instruction.

If the cost overhead is less than or equal to the cost benefit, it means that even if the execution of the execution instruction brings additional cost to the artificial intelligence terminal, it is not conducive to the artificial intelligence terminal, but at the same time can bring equal or more cost to other terminals. income. When the artificial intelligence terminal and other terminals are regarded as a whole, the execution of the execution instruction does not bring additional overhead as a whole, and the overall benefit is realized, and the artificial intelligence terminal selects to execute the execution instruction.

Through the implementation of the embodiment, the artificial intelligence terminal executes execution instructions from other terminals when its own cost overhead is less than or equal to the cost benefit of other terminals. In the process of behavior control, in addition to its own situation, it is also considered. In the case of other terminals, the behavior control based on the altruistic criteria is realized, which makes the behavioral norms of the artificial intelligence terminal more in line with human moral requirements, more reasonable and meets the requirements of users.

As shown in FIG. 2, the second embodiment of the behavior control method of the artificial intelligence terminal of the present invention is based on the first embodiment of the behavior control method of the artificial intelligence terminal of the present invention, and notifies other after deciding whether to execute or not executing the execution instruction. The terminal, and the artificial intelligence terminal calculates the cost benefit by itself. This embodiment is an extension of the first embodiment of the behavior control method of the artificial intelligence terminal of the present invention, and the same portions are not described herein again. This embodiment includes:

S111: The artificial intelligence terminal receives an execution instruction from another terminal.

S112: The artificial intelligence terminal calculates the cost incurred by executing the execution instruction itself.

S113: The artificial intelligence terminal calculates the cost benefit.

The order of execution of this step and step S112 is merely illustrative, and the sequence may be executed or reversed at the same time.

S114: The artificial intelligence terminal determines whether the cost overhead is less than or equal to the cost benefit of other terminals.

If the cost overhead is less than or equal to the cost benefit, then the process goes to step S115; otherwise, the process goes to step S117.

S115: The artificial intelligence terminal executes an execution instruction.

Go to step S116.

S116: The artificial intelligence terminal sends a successful execution notification to other terminals.

A successful execution notification is used to indicate that the execution instruction has been executed. End the process.

S117: The artificial intelligence terminal does not execute the execution instruction.

Go to step S118.

S118: The artificial intelligence terminal sends a rejection execution notification to other terminals.

The refusal execution notification is used to indicate that the execution instruction was not executed. In addition, the rejection notice may also include a reason for rejection. End the process.

As shown in FIG. 3, the third embodiment of the behavior control method of the artificial intelligence terminal of the present invention is based on the first embodiment of the behavior control method of the artificial intelligence terminal of the present invention, and notifies other after deciding to execute or not execute the execution instruction. The terminal, and the artificial intelligence terminal receives cost benefits from other terminals. This embodiment is an extension of the first embodiment of the behavior control method of the artificial intelligence terminal of the present invention, and the same portions are not described herein again. This embodiment includes:

S121: The artificial intelligence terminal receives an execution instruction from another terminal.

S122: The artificial intelligence terminal calculates the cost incurred by executing the execution instruction itself.

S123: The artificial intelligence terminal receives the cost benefit from other terminals.

The order of execution of this step and steps S121 and S122 is merely illustrative, and the sequence may be executed or reversed at the same time.

S124: The artificial intelligence terminal determines whether the cost overhead is less than or equal to the cost benefit of other terminals.

If the cost overhead is less than or equal to the cost benefit, then the process goes to step S125; otherwise, the process goes to step S127.

S125: The artificial intelligence terminal executes an execution instruction.

S126: The artificial intelligence terminal sends a successful execution notification to other terminals.

S127: The artificial intelligence terminal does not execute the execution instruction.

S128: The artificial intelligence terminal sends a rejection execution notification to other terminals.

In the above two embodiments, the artificial intelligence terminal notifies the other terminal whether the execution of the instruction is executed in an explicit manner, that is, by sending a success/rejection execution notification. In other embodiments, the artificial intelligence terminal may also select Partially notified in a hidden way. For example, if a notification time is set, if the artificial intelligence terminal decides to execute the execution instruction, the notification message is sent to the other terminal within the notification time, and if the artificial intelligence terminal decides not to execute the execution instruction, the notification message is not sent. Or conversely, if the artificial intelligence terminal decides not to execute the execution instruction, the notification message is sent to the other terminal within the notification time, and if the artificial intelligence terminal decides to execute the execution instruction, the notification message is not sent.

As shown in FIG. 4, the fourth embodiment of the behavior control method of the artificial intelligence terminal of the present invention includes:

S21: The artificial intelligence terminal receives an execution instruction from another terminal;

The main difference between this embodiment and the first embodiment of the behavior control method of the artificial intelligence terminal of the present invention is that it is based on the relationship between the weighted cost overhead and the weighted cost benefit rather than the relationship between the cost overhead and the cost benefit. Execution instructions are executed, and the same/similar parts are not repeated here.

S22: The artificial intelligence terminal calculates the cost incurred by executing the execution instruction itself.

S23: The artificial intelligence terminal determines whether the weighted cost overhead is less than or equal to the weighted cost benefit of other terminals.

The weighted cost overhead is the product of the weight of the artificial intelligence terminal and the cost overhead. The weighted cost benefit is the product of the weight of other terminals and the cost benefit. The cost benefit is generated by other terminals executing instructions by the artificial intelligence terminal. For example, other terminals are artificial. The intelligent terminal executes execution instructions to reduce computational, motion, power consumption, etc. and/or the benefits of the user of other terminals due to the execution of the execution instructions by the artificial intelligence terminal.

Before this step is performed, the artificial intelligence terminal needs to obtain the weighted cost overhead and the weighted cost benefit. The artificial intelligence terminal can multiply its own weight by the cost overhead to obtain a weighted cost overhead. The artificial intelligence terminal can calculate the cost return by itself and multiply the weight of other terminals to obtain the weighted cost return. It can also receive the cost income sent by other terminals and then multiply the weight of other terminals to obtain the weighted cost return, or directly receive the other terminal. Weighted cost benefit.

In general, the weight of an artificial intelligence terminal/other terminal is determined by its priority. The priority can be determined only by the attributes of the artificial intelligence terminal/other terminal itself. For example, still taking the self-driving car as an example, the priority of the ambulance/rescue/police car can be set to be the highest, and the priority of the school bus/transit bus is second. Then there is the ordinary manned car, which has the lowest priority. In addition, the user attribute of the artificial intelligence terminal can also be considered in the priority setting. For example, for an ordinary manned self-driving car, the priority of the self-driving car at the destination of the airport/train station/car station/school/hospital can be set to be high. For autonomous vehicles that are destined for other locations, the higher the priority of the destination, the higher the priority of the number of passengers.

S24: If the weighted cost overhead is less than or equal to the weighted cost benefit, the artificial intelligence terminal executes the execution instruction.

If the weighted cost overhead is less than or equal to the weighted cost benefit, it means that even if the execution of the execution instruction brings additional weighted cost overhead to the artificial intelligence terminal, it is not conducive to the artificial intelligence terminal, but at the same time can bring equal or more to other terminals. More weighted cost benefits. When the artificial intelligence terminal and other terminals are regarded as a whole, the execution of the execution instruction does not bring additional overhead as a whole, and the overall benefit is realized, and the artificial intelligence terminal selects to execute the execution instruction.

Through the implementation of the embodiment, the artificial intelligence terminal executes execution instructions from other terminals when its weighted cost overhead is less than or equal to the weighted cost benefit of other terminals. In the process of behavior control, in addition to its own situation, Considering the situation of other terminals, the behavior control based on altruistic criteria is realized, which makes the behavioral rules of artificial intelligence terminals more in line with human moral requirements, more reasonable, meets the requirements of users, and takes into account artificial intelligence terminals in the behavior control process. The weight of other terminals further improves the accuracy of the judgment.

As shown in FIG. 5, the fifth embodiment of the behavior control method of the artificial intelligence terminal of the present invention is based on the fourth embodiment of the behavior control method of the artificial intelligence terminal of the present invention, and notifies other after deciding to execute or not execute the execution instruction. The terminal, and the artificial intelligence terminal calculates the cost benefit by itself. This embodiment is an extension of the fourth embodiment of the behavior control method of the artificial intelligence terminal of the present invention, and the same portions are not described herein again. This embodiment includes:

S210: The artificial intelligence terminal receives the weight of itself and other terminals.

The artificial intelligence terminal can receive the weight assigned to itself from the server or the control center, and receive the weights of other terminals from the server, the control center, or other terminals. In other embodiments, the artificial intelligence terminal may locally store the weight of itself and/or other terminals, in which case this step may be partially or completely omitted.

This step only needs to be performed before step S214, and the order of execution between steps S211, S212, and S213 is not limited.

S211: The artificial intelligence terminal receives an execution instruction from another terminal.

S212: The artificial intelligence terminal calculates the cost incurred by executing the execution instruction itself.

S213: The artificial intelligence terminal calculates the cost benefit generated by the other terminal due to the execution of the execution instruction by the artificial intelligence terminal.

The order of execution of this step and step S212 is merely illustrative, and the sequence may be executed or reversed at the same time.

S214: The artificial intelligence terminal calculates a weighted cost overhead and a weighted cost benefit.

S215: The artificial intelligence terminal determines whether the weighted cost overhead is less than or equal to the weighted cost benefit.

If the weighted cost overhead is less than or equal to the weighted cost benefit, then the process goes to step S216; otherwise, the process goes to step S218.

S216: The artificial intelligence terminal executes an execution instruction.

S217: The artificial intelligence terminal sends a successful execution notification to other terminals.

S218: The artificial intelligence terminal does not execute the execution instruction.

S219: The artificial intelligence terminal sends a rejection execution notification to other terminals.

As shown in FIG. 6, the sixth embodiment of the behavior control method of the artificial intelligence terminal of the present invention is based on the fourth embodiment of the behavior control method of the artificial intelligence terminal of the present invention, and notifies other after deciding to execute or not execute the execution instruction. The terminal, and the weighted cost benefit is calculated by the artificial intelligence terminal after receiving the cost benefit. This embodiment is an extension of the fourth embodiment of the behavior control method of the artificial intelligence terminal of the present invention, and the same portions are not described herein again. This embodiment includes:

S220: The artificial intelligence terminal receives the weight of itself and other terminals.

Artificial intelligence can receive weights assigned to itself from the server or control center and receive weights from other terminals from servers, control centers, or other terminals. In other embodiments, the artificial intelligence terminal may locally store the weight of itself and/or other terminals, in which case this step may be partially or completely omitted.

This step only needs to be performed before step S224, and the order of execution between steps S221, S222, and S223 is not limited.

S221: The artificial intelligence terminal receives an execution instruction from another terminal;

S222: The artificial intelligence terminal calculates the cost incurred by executing the execution instruction itself.

S223: The artificial intelligence terminal receives the cost benefit from other terminals.

The execution order of this step and steps S220, S221, and S222 is merely illustrative, and the sequence may be executed or reversed at the same time.

S224: The artificial intelligence terminal calculates a weighted cost overhead and a weighted cost benefit.

S225: The artificial intelligence terminal determines whether the weighted cost overhead is less than or equal to the weighted cost benefit.

If the weighted cost overhead is less than or equal to the weighted cost benefit, then the process goes to step S226; otherwise, the process goes to step S228.

S226: The artificial intelligence terminal executes an execution instruction.

S227: The artificial intelligence terminal sends a successful execution notification to other terminals.

S228: The artificial intelligence terminal does not execute the execution instruction.

S229: The artificial intelligence terminal sends a rejection execution notification to other terminals.

As shown in FIG. 7, the seventh embodiment of the behavior control method of the artificial intelligence terminal of the present invention is based on the fourth embodiment of the behavior control method of the artificial intelligence terminal of the present invention, and notifies other after deciding whether to execute or not executing the execution instruction. The terminal, and the weighted cost benefit is calculated by the artificial intelligence terminal after receiving the cost benefit. This embodiment is an extension of the fourth embodiment of the behavior control method of the artificial intelligence terminal of the present invention, and the same portions are not described herein again. This embodiment includes:

S230: The artificial intelligence terminal receives its own weight.

Artificial intelligence can receive weights assigned to itself from the server or control center. In other embodiments, the artificial intelligence terminal may have its own weight stored locally, in which case this step may be omitted.

This step only needs to be performed before step S234, and the order of execution between steps S231, S232, and S233 is not limited.

S231: The artificial intelligence terminal receives an execution instruction from another terminal;

S232: The artificial intelligence terminal calculates the cost incurred by executing the execution instruction itself.

S233: The artificial intelligence terminal calculates a weighted cost overhead.

S234: The artificial intelligence terminal receives the weighted cost benefit from other terminals.

The order of execution of this step and step S230 to step S233 is merely illustrative, and the sequence may be executed or reversed at the same time.

S235: The artificial intelligence terminal determines whether the weighted cost overhead is less than or equal to the weighted cost benefit.

If the weighted cost overhead is less than or equal to the weighted cost benefit, then the process goes to step S236; otherwise, the process goes to step S238.

S236: The artificial intelligence terminal executes an execution instruction.

S237: The artificial intelligence terminal sends a successful execution notification to other terminals.

S238: The artificial intelligence terminal does not execute the execution instruction.

S239: The artificial intelligence terminal sends a rejection execution notification to other terminals.

In the foregoing three embodiments, the artificial intelligence terminal notifies the other terminal whether the execution of the instruction is executed in an explicit manner, that is, by sending a success/rejection execution notification. In other embodiments, the artificial intelligence terminal may also select Partially notified in a hidden way. For example, if a notification time is set, if the artificial intelligence terminal decides to execute the execution instruction, the notification message is sent to the other terminal within the notification time, and if the artificial intelligence terminal decides not to execute the execution instruction, the notification message is not sent. Or conversely, if the artificial intelligence terminal decides not to execute the execution instruction, the notification message is sent to the other terminal within the notification time, and if the artificial intelligence terminal decides to execute the execution instruction, the notification message is not sent.

As shown in FIG. 8, the eighth embodiment of the behavior control method of the artificial intelligence terminal of the present invention includes:

S31: The artificial intelligence terminal formulates an execution strategy.

The execution policy may include an instruction to be executed by the artificial intelligence terminal itself and/or a number of related instructions to be executed by other terminals, and the two instructions generally cooperate with each other.

The artificial intelligence terminal may be an intelligent robot, an autonomous driving vehicle (for example, an autonomous driving vehicle) or other terminal having data analysis processing capability. Communication connection between other terminals and the terminal. Other terminals may be artificial intelligence terminals or non-intelligent terminals.

Still using the self-driving car as an example, the artificial intelligence terminal is the self-driving car C, the other terminals include the self-driving cars D and E, the three are driving in the same lane and the D and E are in front of the C, and the execution strategy formulated by C includes itself. Straight overtaking, D and E temporarily move to the side lane.

S32: The artificial intelligence terminal separately calculates the cost incurred by the other terminals involved in the execution strategy due to executing the execution strategy and the cost benefits generated by the artificial intelligence terminal itself by executing the execution strategy by other terminals.

The cost overhead may refer to an increase in computation, motion, energy consumption, and the like of other terminals and/or their users due to execution of the execution policy by other terminals. The cost benefit may be the computational, motion, energy, and the like that the artificial intelligence terminal and/or its user reduces due to the execution of the execution strategy by the artificial intelligence terminal.

Still using the above example, the cost overhead for D and E is the additional cost of moving from the current lane to the side lane and then from the side lane back to the current lane; for C, if D and If E does not cooperate with the implementation of the execution strategy, then C needs to move from the current lane to the side lane beyond D and E and then move back from the side lane to the current lane. If D and E cooperate to execute the execution strategy, then C only needs to travel straight. Yes, the difference between the two is the cost benefit of C.

S33: The artificial intelligence terminal judges whether to execute the execution strategy according to the overall benefit principle according to the cost benefit and the cost overhead.

The criterion for judgment is the overall benefit, that is, the implementation of the execution strategy does not bring additional overhead as a whole, and the whole includes artificial intelligence terminals and all other terminals. Specifically, the artificial intelligence terminal may execute the execution strategy if the cost benefit is greater than or equal to the sum of the cost expenses, or execute the execution strategy if the weighted cost benefit is greater than or equal to the sum of the weighted cost expenses.

In the case that the overall benefit criterion is met, the artificial intelligence terminal may execute the execution policy by itself and/or notify other terminals involved in the execution of the execution policy.

Through the implementation of the embodiment, the artificial intelligence terminal evaluates the cost benefit and the cost overhead of the execution strategy according to the overall benefit criterion after formulating the execution strategy, thereby determining whether to execute the execution strategy, in addition to itself in the process of controlling the behavior. In addition to the situation of other terminals, the behavior control based on altruistic criteria is also implemented, so that the behavioral rules of artificial intelligence terminals are more in line with human moral requirements, more reasonable, meet the requirements of users, and formulate implementation strategies. Behavioral control is implemented to avoid unnecessary overhead caused by inappropriate execution policies to themselves and/or other terminals.

As shown in FIG. 9, the ninth embodiment of the behavior control method of the artificial intelligence terminal of the present invention is based on the eighth embodiment of the behavior control method of the artificial intelligence terminal of the present invention, and judges whether to execute according to the cost benefit and the cost overhead itself. Execution strategy. This embodiment is an extension of the eighth embodiment of the behavior control method of the artificial intelligence terminal of the present invention, and the same portions are not described herein again. This embodiment includes:

S311: The artificial intelligence terminal formulates an execution strategy.

S312: The artificial intelligence terminal separately calculates the cost incurred by the other terminals involved in the execution strategy due to executing the execution strategy and the cost benefits generated by the artificial intelligence terminal itself by executing the execution strategy by other terminals.

S313: The artificial intelligence terminal determines whether the cost benefit is greater than or equal to the sum of the cost expenses.

If the total number of other terminals is 1, the sum of the cost overheads is equal to the cost overhead of the other other terminals.

If the cost benefit is greater than or equal to the sum of the cost, it means that even if the implementation of the execution strategy brings additional cost to other terminals, it is not conducive to other terminals, but at the same time can bring equal or more to the artificial intelligence terminal. The cost benefit, satisfying the overall benefit principle, jumps to step S314; otherwise, it jumps to step S315. In other embodiments, if the cost benefit is less than the sum of the cost overheads, the process can be terminated.

S314: The artificial intelligence terminal itself performs an execution policy and/or notifies other terminals to execute an execution policy.

End the process.

S315: Modify the execution strategy.

Then it jumps to step S312.

As shown in FIG. 10, the tenth embodiment of the behavior control method of the artificial intelligence terminal of the present invention is based on the weighted cost benefit and the weighted cost overhead of the eighth embodiment of the behavior control method of the artificial intelligence terminal of the present invention. Execute the execution strategy. This embodiment is an extension of the eighth embodiment of the behavior control method of the artificial intelligence terminal of the present invention, and the same portions are not described herein again. This embodiment includes:

S321: The artificial intelligence terminal formulates an execution strategy.

S322: The artificial intelligence terminal separately calculates the cost incurred by the other terminals involved in the execution strategy due to executing the execution strategy, and the cost benefits generated by the artificial intelligence terminal itself by executing the execution strategy by other terminals.

S323: The artificial intelligence terminal calculates the weighted cost overhead and the weighted cost benefit according to the cost overhead and the cost return respectively.

The weighted cost benefit is the product of the cost benefit and the weight of the artificial intelligence terminal, the weighted cost overhead is the weighted sum of the cost overheads of all other terminals, and the weight in the weighted sum is the weight of each other terminal. If the total number of other terminals is 1, the sum of the weighted cost overheads is equal to the weighted cost overhead of the unique other terminal.

In general, the weight of an artificial intelligence terminal/other terminal is determined by its priority. The priority can be determined only by the attributes of the artificial intelligence terminal itself. For example, still taking the self-driving car as an example, the ambulance/firetruck/police car can be set to have the highest priority, the school bus/transit bus priority is followed, and then the ordinary For manned vehicles, trucks have the lowest priority. In addition, the user attribute of the artificial intelligence terminal can also be considered in the priority setting. For example, for an ordinary manned self-driving car, the priority of the self-driving car at the destination of the airport/train station/car station/school/hospital can be set to be high. For autonomous vehicles that are destined for other locations, the higher the priority of the destination, the higher the priority of the number of passengers.

S324: The artificial intelligence terminal determines whether the weighted cost benefit is greater than or equal to the sum of the weighted cost expenses.

If the weighted cost benefit is greater than or equal to the sum of the weighted cost expenses, it means that even if the execution of the execution strategy brings additional weighted cost overhead to other terminals, it is not conducive to other terminals, but at the same time can bring equality to the artificial intelligence terminal. Or more weighted cost benefits, satisfying the overall benefit principle, jump to step S325; otherwise, jump to step S326. In other embodiments, the process may end if the weighted cost benefit is less than the sum of the weighted cost overheads.

S325: The artificial intelligence terminal itself performs an execution policy and/or notifies other terminals to execute an execution policy.

End the process.

S326: Modify the execution strategy.

Then it jumps to step S322.

Compared with the previous embodiment, the present embodiment adopts a weighted cost benefit/overhead instead of a cost benefit/overhead in the judging process, and further improves the accuracy of the judging by considering the weights of the artificial intelligence terminal and other terminals.

As shown in FIG. 11, the eleventh embodiment of the behavior control method of the artificial intelligence terminal of the present invention includes:

S41: The artificial intelligence terminal accumulates the self-cost return generated by the other terminal to execute its own instruction, obtains the accumulated cost of the own cost, and accumulates other people's cost benefits generated by the other terminal's execution of the instruction of the other terminal by the artificial intelligence terminal to obtain others. The cumulative value of cost and revenue.

The artificial intelligence terminal may be an intelligent robot, an autonomous driving vehicle (for example, an autonomous driving vehicle) or other terminal having data analysis processing capability. Communication connection between other terminals and the terminal. Other terminals may be artificial intelligence terminals or non-intelligent terminals. The execution instructions of the artificial intelligence terminal/other terminal may be given by the artificial intelligence terminal/other terminal, or may be from the user, such as an artificial intelligence terminal/other terminal from its input device (touch screen, keyboard, microphone, camera, The mouse, etc.) receives the instruction given by the user.

The accumulated value of the cost-earnings of itself may be the sum or weighted sum of the self-costs and returns within the preset time period, and the accumulated value of the cost-earnings of others may be the sum or weighted sum of the cost-of-interests of others within the preset time period.

If it is a weighted sum, the artificial intelligence terminal needs to obtain the weight of itself and/or other terminals before this step. The artificial intelligence terminal can receive the weight assigned to itself from the server or the control center, and receive the weights of other terminals from the server, the control center, or other terminals. Alternatively, the artificial intelligence terminal can directly read the weights of the locally saved itself and/or other terminals, and then receive the local unsaved weights.

S42: The artificial intelligence terminal controls its own decision according to the accumulated cost value of itself and the accumulated value of the cost of others, so that the difference between the accumulated value of the cost-earnings and the accumulated value of the cost-earnings of others is kept within a preset range.

A decision can refer to an execution instruction that determines whether to execute itself and/or other terminals. The difference between the accumulated cost of the own cost and the accumulated value of the cost of the other person may refer to the difference between the two, or the absolute value of the difference between the two.

Through the implementation of the embodiment, the artificial intelligence terminal controls its own decision so that the difference between the accumulated cost of the cost and the accumulated value of the cost of the other person is kept within the preset range, and in the process of behavior control, the self and others are guaranteed. The balance of the accumulated value of the income realizes the behavior control based on the altruistic criteria rather than the self-interest, so that the behavioral norms of the artificial intelligence terminal are more in line with the human moral requirements, more reasonable and meet the requirements of the users.

As shown in FIG. 12, the twelfth embodiment of the behavior control method of the artificial intelligence terminal of the present invention is based on the eleventh embodiment of the behavior control method of the artificial intelligence terminal of the present invention, and S42 includes:

S421: The artificial intelligence terminal calculates the self-cost return and/or the cost benefit of others generated by executing the instruction.

Execution instructions may include their own execution instructions and/or execution instructions of other terminals.

S422: The artificial intelligence terminal updates the accumulated cost value of the own cost and/or the accumulated value of the cost of the other person according to the calculation result.

S423: The artificial intelligence terminal determines whether the difference between the updated self-cost return cumulative value and the other person's cost-revenue accumulated value belongs to a preset range.

If yes, the process goes to step S424; otherwise, the process goes to step S425.

S424: Execute an execution instruction.

If the execution instruction includes its own execution instruction, the artificial intelligence terminal may execute the own execution instruction itself and/or notify the relevant other terminal to execute the own execution instruction. If the execution instruction includes the execution instruction of the other terminal, the artificial intelligence terminal may execute the execution instruction of the other terminal. After the execution is completed, the artificial intelligence terminal may further select to notify the corresponding other terminal in an explicit or implicit manner, which may be specifically referred to. Corresponding description of the second and third embodiments of the behavior control method of the artificial intelligence terminal of the present invention.

S425: Execution instruction is not executed.

If the execution instruction includes the execution instruction of the other terminal, the artificial intelligence terminal may select to notify other terminals in an explicit or implicit manner after the step, and specifically refer to the second and third implementations of the behavior control method of the artificial intelligence terminal of the present invention. Corresponding description of the example.

As shown in FIG. 13, the thirteenth embodiment of the behavior control method of the artificial intelligence terminal of the present invention includes:

S51: The artificial intelligence terminal formulates at least two candidate execution strategies.

The artificial intelligence terminal may be an intelligent robot, an autonomous driving vehicle (for example, an autonomous driving vehicle) or other terminal having data analysis processing capability. In this embodiment, the artificial intelligence terminal is used as a formulator of the execution strategy.

Each candidate execution strategy may include candidate instructions of the artificial intelligence terminal itself and/or candidate instructions of several related other terminals, and the two instructions generally cooperate with each other.

Still using the self-driving car as an example, the artificial intelligence terminal is the self-driving car F, the other terminals include the self-driving cars G and H, the three are driving in the same lane and the G and H are in front of the F, the first candidate formulated by F The execution strategy includes its own straight-through overtaking, G and H temporarily move to the side lane, and the second candidate execution strategy includes self-overtaking from the side lane, and G and H maintain straight-line driving.

S52: The artificial intelligence terminal sends at least two candidate execution policies to other terminals, so that other terminals respectively evaluate the executable degree of each candidate execution policy.

Communication connection between other terminals and the terminal. Other terminals can follow the principle of complete self-interest in judging the enforceability, and can also follow the altruistic principle in consideration of the situation of the artificial intelligence terminal.

The value of the executable may be only a or b, where a means executable, b means unexecutable, and a>b, such as a=1, b=0. The artificial intelligence terminal can limit each other terminal to a maximum/minimum/only selectable (for example one) executable candidate execution strategy, or can be judged by other terminals.

The value of the executable degree may also be a plurality of consecutive or discontinuous numbers, and the greater the value of the executable degree, the stronger the willingness of other terminals to execute their corresponding candidate execution strategies.

The other terminals involved in the different candidate execution policies may be different. The artificial intelligence terminal may send the candidate execution policy only to other terminals involved in each candidate execution policy, or may uniformly send to all other terminals involved in all candidate execution policies. All candidate execution strategies.

S53: The artificial intelligence terminal receives the executable degree from other terminals.

S54: The artificial intelligence terminal selects an optimal execution policy from the at least two candidate execution strategies according to the executable degree.

For each candidate execution policy, the artificial intelligence terminal may obtain the comprehensive executable degree of the candidate execution strategy according to the executable degree statistics of the candidate execution policy by each other terminal, and then select according to the comprehensive executable degree of all candidate execution strategies. . The calculation process of the candidate execution strategy evaluation by the artificial intelligence terminal itself may or may not be considered in the calculation process of the comprehensive executable degree of each candidate execution strategy.

S55: The artificial intelligence terminal performs an optimal execution strategy.

Specifically, the artificial intelligence terminal may perform an optimal execution policy by itself and/or notify other terminals to perform an optimal execution strategy.

Through the implementation of the embodiment, the artificial intelligence terminal selects an optimal execution strategy and performs an optimal execution strategy according to the executable degree obtained by the other terminal for at least two candidate execution policy evaluations, in addition to itself in the process of controlling the behavior. In addition to the situation of other terminals, the behavior control based on altruistic criteria is also implemented, so that the behavioral guidelines of artificial intelligence terminals are more in line with human moral requirements, more reasonable, meet the requirements of users, and formulate implementation strategies. The process comprehensively considers the executable degree obtained by other terminal evaluations, optimizes the execution strategy of the final execution, and avoids the unnecessary overhead caused by the inappropriate execution strategy to itself and/or other terminals.

As shown in FIG. 14, the fourteenth embodiment of the behavior control method of the artificial intelligence terminal of the present invention is based on the thirteenth embodiment of the behavior control method of the artificial intelligence terminal of the present invention, and S54 includes:

S541: The artificial intelligence terminal counts the comprehensive executable degree of each candidate execution strategy.

The comprehensive enforceability of each candidate execution strategy is the sum or weighted sum of its executables, and if it is a weighted sum, the weight of the weighted sum is the weight of other terminals.

Still following the above example, the weight of the self-driving car G is sg, the executables for the first and second candidate execution strategies are g1 and g2, respectively; the weight of the self-driving car H is sh, The enforceability of the first and second candidate execution strategy evaluations is h1 and h2, respectively. Then the comprehensive executable degree of the first candidate execution strategy is g1+h1 (sum) or sg*g1+sh*h1 (weighted sum), and the comprehensive executable degree of the second candidate execution strategy is g2+h2 (sum) Or sg*g2+sh*h2 (weighted sum).

S542: The artificial intelligence terminal selects the candidate execution strategy with the largest comprehensive executable as the best execution strategy.

As shown in FIG. 15, the fifteenth embodiment of the behavior control method of the artificial intelligence terminal of the present invention includes:

S61: The artificial intelligence terminal receives at least one candidate execution policy from another terminal.

The artificial intelligence terminal may be an intelligent robot, an autonomous driving vehicle (for example, an autonomous driving vehicle) or other terminal having data analysis processing capability. In this embodiment, the artificial intelligence terminal is used as an evaluation party of the execution strategy.

S62: The artificial intelligence terminal judges the executable degree of each candidate execution strategy.

For each candidate execution strategy, the artificial intelligence terminal can refer to any one of the first to seventh, eleventh and twelfth embodiments of the behavior control method of the artificial intelligence terminal of the present invention and the method provided by the combination of non-collision Evaluation, if it is determined that the candidate execution strategy can be executed, it is determined that the value of its executable degree is a, otherwise the value of its executable degree is determined to be b, a>b. Of course, other methods can also be used for judging.

S63: The artificial intelligence terminal sends the executable degree to other terminals, so that the other terminals select the best execution policy from the at least two candidate execution policies according to the executable degree and execute the optimal execution strategy.

For details, refer to the descriptions of the thirteenth and fourteenth embodiments of the behavior control method of the artificial intelligence terminal of the present invention.

As shown in FIG. 16, the sixteenth embodiment of the behavior control method of the artificial intelligence terminal of the present invention is based on the fifteenth embodiment of the behavior control method of the artificial intelligence terminal of the present invention, and S62 includes:

S621: The artificial intelligence terminal separately calculates the cost cost generated by each candidate execution strategy and the cost benefit generated by the other artificial terminal by executing each candidate execution strategy.

S622: The artificial intelligence terminal calculates the executable degree according to the cost overhead and the cost benefit.

In general, the formula for calculating the enforceability should satisfy the altruistic principle. For example, for a candidate execution strategy, the artificial intelligence terminal calculates a cost cost of x, a cost benefit of y, and an executable degree of (yx)/(y+x), (yx)/y, or (yx)/ x and so on.

As shown in FIG. 17, the first embodiment of the artificial intelligence terminal of the present invention includes a processor 110 and a communication circuit 120, and the processor 110 is connected to the communication circuit 120.

The communication circuit 120 is used for transmitting and receiving data, and is an interface for the artificial intelligence terminal to communicate with other terminals.

The processor 110 controls the operation of the artificial intelligence terminal, and the processor 110 may also be referred to as a CPU (Central Processing). Unit, central processing unit). Processor 110 may be an integrated circuit chip with signal processing capabilities. The processor 110 can also be a general purpose processor, a digital signal processor (DSP), an application specific integrated circuit (ASIC), an off-the-shelf programmable gate array (FPGA) or other programmable logic device, a discrete gate or transistor logic device, and discrete hardware components. . The general purpose processor may be a microprocessor or the processor or any conventional processor or the like.

The processor 110 is operative to execute instructions to implement any of the embodiments of the behavioral control method of the artificial intelligence terminal of the present invention and any non-conflicting combination.

As shown in FIG. 18, the first embodiment of the computer storage medium of the present invention includes a memory 200 in which a program is stored, the program can be executed to implement any embodiment of the behavior control method of the artificial intelligence terminal of the present invention, and any non-conflicting Combine the methods provided.

The memory 200 can include a read only memory (ROM, Read-Only) Memory), Random Access Memory (RAM), Flash Memory, hard disk, optical disk, etc.

In the several embodiments provided by the present invention, it should be understood that the disclosed artificial intelligence terminal can be implemented in other manners. For example, as described above The device implementation is only illustrative. For example, the division of the module or resource unit is only a logical function division. In actual implementation, there may be another division manner. For example, multiple resource units or components may be combined or integrated. Go to another system, or some features can be ignored or not executed. In addition, the mutual coupling or direct coupling or communication connection shown or discussed may be an indirect coupling or communication connection through some interface, device or resource unit, and may be in an electrical, mechanical or other form.

The resource units described as separate components may or may not be physically separated. The components displayed as resource units may or may not be physical resource units, that is, may be located in one place, or may be distributed to multiple networks. On the resource unit. Some or all of the resource units may be selected according to actual needs to achieve the purpose of the solution of the present embodiment.

In addition, each functional resource unit in each embodiment of the present invention may be integrated into one processing resource unit, or each resource unit may exist physically separately, or two or more resource units may be integrated into one resource unit. The above integrated resource unit can be implemented in the form of hardware or in the form of a software function resource unit.

The integrated resource unit, if implemented in the form of a software functional resource unit and sold or used as a standalone product, may be stored in a computer readable storage medium. Based on such understanding, the technical solution of the present invention, which is essential or contributes to the prior art, or all or part of the technical solution, may be embodied in the form of a software product stored in a storage medium. A number of instructions are included to cause a computer device (which may be a personal computer, server, or network device, etc.) or a processor to perform all or part of the steps of the methods of the various embodiments of the present invention. The foregoing storage medium includes: a U disk, a mobile hard disk, a read only memory (ROM, Read-Only) Memory, random access memory (RAM), disk or optical disk, and other media that can store program code.

The above is only the embodiment of the present invention, and is not intended to limit the scope of the invention, and the equivalent structure or equivalent process transformations made by the description of the invention and the drawings are directly or indirectly applied to other related technologies. The fields are all included in the scope of patent protection of the present invention.

Claims

A behavior control method for an artificial intelligence terminal, comprising:

The artificial intelligence terminal formulates at least two candidate execution strategies;

The artificial intelligence terminal sends the at least two candidate execution policies to other terminals, so that the other terminals respectively evaluate the executable degree of each of the candidate execution policies;

The artificial intelligence terminal receives the executable degree from the other terminal;

The artificial intelligence terminal selects an optimal execution policy from the at least two candidate execution strategies according to the executable degree;

The artificial intelligence terminal performs the best execution strategy.
The method according to claim 1, wherein the artificial intelligence terminal selects the best execution policy from the candidate execution policies according to the executable degree, including:

The artificial intelligence terminal counts the comprehensive executable degree of each of the candidate execution strategies;

The artificial intelligence terminal selects the candidate execution strategy with the largest comprehensive executable as the best execution strategy.
The method according to claim 2, wherein the integrated executable degree of each of the candidate execution strategies is a sum or a weighted sum of the executable degrees, and the weights of the weighted sums are Weights.
The method according to any one of claims 1 to 3, wherein the performing the optimal execution strategy by the artificial intelligence terminal comprises:

The artificial intelligence terminal itself performs the best execution policy and/or notifies the other terminal to execute the best execution policy.
A behavior control method for an artificial intelligence terminal, comprising:

The artificial intelligence terminal receives at least one candidate execution policy from other terminals;

The artificial intelligence terminal separately evaluates the executable degree of each of the candidate execution strategies;

The artificial intelligence terminal sends the executable degree to the other terminal, so that the other terminal selects an optimal execution policy from the at least two candidate execution policies according to the executable degree and executes the most Good execution strategy.
The method of claim 5 wherein:

The determining, by the artificial intelligence terminal, the executable degree of each of the candidate execution strategies includes:

The artificial intelligence terminal separately calculates a cost cost generated by each of the candidate execution strategies and a cost benefit generated by the other terminal by the artificial intelligence terminal to execute each of the candidate execution strategies;

The artificial intelligence terminal determines the executable degree of each of the candidate execution policies according to the cost overhead and the cost benefit of each of the candidate execution policies.
The method of claim 6 wherein:

Determining, by the artificial intelligence terminal, the executable degree of each of the candidate execution policies according to the cost cost and the cost benefit of each of the candidate execution policies comprises:

Determining, by the artificial intelligence terminal, whether the cost cost of each of the candidate execution policies is less than or equal to the cost benefit;

If the cost overhead is less than or equal to the cost benefit, the artificial intelligence terminal determines that the executable value is a, otherwise determines that the executable value has a value of b, a>b.
The method of claim 6 wherein:

Determining, by the artificial intelligence terminal, the executable degree of each of the candidate execution policies according to the cost cost and the cost benefit of each of the candidate execution policies comprises:

Determining, by the artificial intelligence terminal, whether a weighted cost overhead of each of the candidate execution policies is less than or equal to a weighted cost benefit, where the weighted cost overhead is a product of the cost overhead and a weight of the artificial intelligence terminal, the weighting The cost benefit is the product of the cost benefit and the weight of the other terminal;

If the weighted cost overhead is less than or equal to the weighted cost benefit, the artificial intelligence terminal determines that the value of the executable degree is a, and otherwise determines that the value of the executable degree is b, a>b.
The method of claim 6 wherein:

Determining, by the artificial intelligence terminal, the executable degree of each of the candidate execution policies according to the cost cost and the cost benefit of each of the candidate execution policies comprises:

The artificial intelligence terminal calculates the executable degree according to the cost overhead and the cost benefit.
The method of claim 6 wherein:

Determining, by the artificial intelligence terminal, the executable degree of each of the candidate execution policies according to the cost cost and the cost benefit of each of the candidate execution policies comprises:

The artificial intelligence terminal calculates the executable degree according to the weighted cost overhead and the weighted cost benefit, where the weighted cost overhead is a product of the cost overhead and the weight of the artificial intelligence terminal, and the weighted cost benefit is The product of the cost benefit and the weight of the other terminals.
The method of claim 5 wherein:

The determining, by the artificial intelligence terminal, the executable degree of each of the candidate execution strategies includes:

And the artificial intelligence terminal accumulates the self-cost return generated by the other terminal to execute the self-instruction, and obtains the self-cost-revenue integrated value, and generates the result that the other terminal executes the instruction of the other terminal by the artificial intelligence terminal. The cost-benefit of others is accumulated to obtain the cumulative value of the cost-earnings of others;

The artificial intelligence terminal separately calculates its own cost benefit generated by executing each of the candidate execution strategies and/or the other person's cost benefit generated by the other intelligent terminal by executing each of the candidate execution strategies;

The artificial intelligence terminal updates the self-cost revenue cumulative value and/or the other person cost revenue cumulative value according to each calculation result;

Determining, by the artificial intelligence terminal, whether the difference between the updated self-cost return cumulative value and the other person's cost-benefit integrated value belongs to a preset range;

If yes, the artificial intelligence terminal determines that the value of the executable degree is a, otherwise determines that the value of the executable degree is b, a>b. .
The method according to claim 11, wherein the cumulative value of the self-cost revenue is a sum of all the self-costs and revenues in the preset time period, and the accumulated value of the cost-others of the others is all in the preset time period. The sum of the cost-benefit of others.
The method according to claim 11, wherein the self-cost-earnings cumulative value is a weighted sum of all of the self-costs and returns within a preset time period, and the other person's cost-benefit cumulative value is all within the preset time period. The weighted sum of the cost benefits of others.
An artificial intelligence terminal, comprising: a processor and a communication circuit, wherein the processor is connected to the communication circuit;

The processor is operative to execute instructions to implement the method of any of claims 1-13.
The terminal according to claim 14, wherein

The artificial intelligence terminal is an intelligent robot or an autonomous driving vehicle.
A computer storage medium, characterized in that the computer storage medium stores a program, the program being executable to implement the behavior control method of the artificial intelligence terminal according to any one of claims 1-13.