Assistance System Trajectory Selection for Smooth Control Transitions

Resolve Bottlenecks,
Find Innovative Solutions
Generate Solutions

Solution Overview

Problem

Technical systems manually controlled by users can enter invalid or unwanted states due to incorrect operations, leading to jerky behavior when assistance systems take over control, reducing user intervention options and potentially causing system misuse.

Innovation Solution

An assistance system that reads user-manipulated variables, generates and evaluates trajectories, and selects control signals based on permissibility and distance metrics to ensure smooth transitions and residual user control, minimizing jerky behavior upon return to user control.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Reliability

If the assistance system takes over control when reaching critical conditions, then system safety is improved, but user intervention options are reduced and jerky behavior occurs upon return of control

Engineering Contradiction:
Improvesystem safetyVSAvoiduser intervention options
Core Design Contradiction:
ReliabilityVSEase of operation

Solution Approach 1:

The assistance system applies partial control by only intervening when critical conditions are detected, while leaving normal operation under user control. The system generates multiple control variants and selects the most appropriate one, applying just enough automated control to ensure safety while preserving maximum user intervention options during non-critical phases.

Inventive Principle:
Principle #16Partial or excessive action

Solution Approach 2:

The system changes control parameters dynamically by adjusting the degree of assistance based on the criticality level. When conditions are not critical, the user maintains full control. When critical conditions are detected, the system transitions to automated control with modified parameters, and during return to user control, it gradually adjusts parameters to avoid jerky behavior.

Inventive Principle:
Principle #35Parameter changes

2Reliability

If the assistance system takes over control completely, then system safety is improved, but jerky behavior occurs upon return of control to the user

Engineering Contradiction:
Improvesystem safetyVSAvoidsystem behavior smoothness
Core Design Contradiction:
ReliabilityVSStability of the object's composition

Solution Approach 1:

The system dynamically adjusts the control mode based on real-time condition assessment. During the transition from automated back to manual control, the system dynamically modifies control parameters and gradually transfers authority, ensuring smooth behavioral changes rather than abrupt switches that would cause jerky system responses.

Inventive Principle:
Principle #15Dynamics

Solution Approach 2:

Before fully returning control to the user, the system performs preliminary actions by gradually reducing automated intervention and preparing the transition. This preliminary phase allows the system to stabilize behavior patterns and prevent jerky responses when control is completely handed back to the user.

Inventive Principle:
Principle #10Preliminary action

3Measurement precision

If multiple control variants are generated and evaluated, then control precision is improved, but computing effort increases

Engineering Contradiction:
Improvecontrol precisionVSAvoidcomputing effort
Core Design Contradiction:
Measurement precisionVSUse of energy by stationary object

Solution Approach 1:

The control variant evaluation process is segmented into hierarchical stages: first evaluating basic safety criteria, then progressively assessing more complex parameters only for variants that pass earlier stages. This segmentation allows the system to generate multiple control variants with high precision while reducing overall computing effort by eliminating unpromising variants early in the evaluation process.

Inventive Principle:
Principle #1Segmentation

Data Source

PatentEP3298462B1Method and assistance system for controlling a technical system
Publication Date: 2021.10.27 SIEMENS AG
  • EP3298462B1 patent drawingFigure 1

AI summary

According to the invention, in order to control a technical system (TS), a user control variable (BS) is read in and a plurality of control variable variants (SV1,..., SVN) of the user control variable (BS) are generated. A respective trajectory (T0, T1,..., TN) of the technical system is extrapolated for the user control variable (BS) and for the control variable variants (SV1,..., SVN), for which a respective reliability is evaluated. Furthermore, a respective distance of each control variable variant (SV1,..., SVN) to the user control variable (BS) is determined. The user control variable (BS) is then selected as a control signal (SIG) for the technical system (TS) in the event that the trajectory (T0, T1,..., TN) extrapolated for the user control variable (BS) is evaluated as reliable. Otherwise, a control variable variant with an extrapolated trajectory (T0, T1,..., TN) evaluated as reliable is selected from the control variable variants as a control signal (SIG), wherein a control variable variant with a low distance is preferably selected. Finally, the control signal (SIG) for controlling the technical system (TS) is emitted.