Hybrid State Machine for Efficient Operation Control

Resolve Bottlenecks,
Find Innovative Solutions
Generate Solutions

Solution Overview

Problem

Current finite state machines and behavior trees face inefficiencies in controlling operations, particularly with numerous states, leading to increased designer workload and difficulty in maintenance, and are not intuitive for handling abnormal state changes.

Innovation Solution

A hybrid state machine is generated using configuration information that integrates state nodes and behavior tree nodes, allowing the server to determine whether to perform operations based on state nodes or behavior tree nodes, thereby improving efficiency and handling abnormal cases.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Ease of operation

If a finite state machine is used to control an individual to perform an operation, then the operation can be controlled through state transitions, but when there are many states, the finite state machine needs to process a large quantity of state conversion relationships, increasing designer workload and maintenance difficulty, and reducing operational efficiency

Engineering Contradiction:
Improveease of controlling individual operationVSAvoidcomplexity of state conversion relationships
Core Design Contradiction:
Ease of operationVSDevice complexity

Solution Approach 1:

The patent combines finite state machine and behavior tree into a hybrid state machine. The finite state machine handles state transitions while the behavior tree manages complex behavior logic within each state. This merging allows the system to process complex operations without proportionally increasing the number of state conversion relationships, thereby reducing designer workload and maintenance difficulty while maintaining operational efficiency.

Inventive Principle:
Principle #5Merging (Combining)

2Productivity

If a behavior tree is used to control an individual to perform an operation, then a series of behaviors can be executed as planned, but the behavior tree is not intuitive in terms of state change and cannot effectively process state changes in abnormal and special cases

Engineering Contradiction:
Improveefficiency of executing planned behaviorsVSAvoidintuitiveness of state change
Core Design Contradiction:
ProductivityVSEase of operation

Solution Approach 1:

The hybrid state machine segments the control system into two distinct parts: the finite state machine component that handles state transitions and abnormal cases (providing intuitive state change tracking), and the behavior tree component that executes planned behavior sequences (ensuring efficient productivity). This segmentation allows each component to excel at its specific function while working together in a unified framework.

Inventive Principle:
Principle #1Segmentation

3Adaptability or versatility

If a finite state machine with many states is implemented, then comprehensive operation control is achieved, but the workload of the designer increases and maintenance becomes inconvenient

Engineering Contradiction:
Improvecomprehensiveness of operation controlVSAvoidease of designing and maintaining state machine
Core Design Contradiction:
Adaptability or versatilityVSEase of manufacture

Solution Approach 1:

The behavior tree acts as an intermediary layer between the finite state machine and the actual operation execution. Instead of managing all operation logic through numerous state transitions, the system uses the behavior tree to handle complex behavior sequences within each state. This intermediary approach maintains comprehensive operation control while significantly reducing the complexity of state conversion relationships, making the system easier to design and maintain.

Inventive Principle:
Principle #24Intermediary (Mediator)

Data Source

PatentEP3425849B1State control method and device
Publication Date: 2024.03.06 TENCENT TECHNOLOGY (SHENZHEN) CO LTD
  • EP3425849B1 patent drawingFigure 1~2
  • EP3425849B1 patent drawingFigure 3~4a
  • EP3425849B1 patent drawingFigure 4b~5

AI summary

The present disclosure discloses a state control method and apparatus. The method includes: detecting, by a server, an operation execution event, the operation execution event being used for requesting to operate a first object in a first application; determining, by the server, whether a next node of a current node that is of the first object and that is in a hybrid state machine is a first state node or a first behavior tree node, nodes in the hybrid state machine including a state node and a behavior tree node, and there being a state change relationship between the nodes in the hybrid state machine; and if determining that the next node of the current node is the first state node, performing, by the server, an operation on the first object, the operation being in a first state corresponding to the first state node; or if determining that the next node of the current state is the first behavior tree node, performing, by the server, an operation on the first object, the operation being in a first behavior tree corresponding to the first behavior tree node. The present disclosure resolves a technical problem in a related technology that efficiency of using a finite state machine to control an individual in an application to perform an operation is low.