AI-Calibrated IMU Positioning for Surgical Navigation Accuracy

Resolve Bottlenecks,
Find Innovative Solutions
Generate Solutions

Solution Overview

Problem

Inertial measurement units (IMUs) used in medical navigation suffer from significant errors in position and orientation estimation due to integration of acceleration data, which amplifies measurement errors over time, and are unsuitable for surgical environments due to cost, size, and ergonomic limitations.

Innovation Solution

A method using an AI system trained with motion data from IMUs, correlated with optical tracking, to predict precise position and orientation without external tracking systems, by capturing motion data along predefined trajectories and creating a linked dataset for neural network training.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Ease of operation

If inertial measurement units are used for navigation in medical technology, then visual obstructions are avoided and ergonomics are improved, but position determination accuracy deteriorates due to error amplification through double integration

Engineering Contradiction:
ImproveergonomicsVSAvoidposition determination accuracy
Core Design Contradiction:
Ease of operationVSMeasurement precision

Solution Approach 1:

The patent introduces an optical tracking system as an intermediary to provide accurate position and orientation data that serves as ground truth for training an AI system. This intermediary system enables the creation of a trained AI model that can later predict position and orientation from acceleration data alone, resolving the accuracy issue while maintaining the ergonomic benefits of IMU usage.

Inventive Principle:
Principle #24Intermediary (Mediator)

Solution Approach 2:

The patent applies preliminary action by training the AI system in advance using datasets generated from optical tracking. This pre-training phase allows the IMU system to achieve high accuracy without requiring real-time optical tracking during actual surgical procedures, thus maintaining ergonomics while improving measurement precision.

Inventive Principle:
Principle #10Preliminary action

2Measurement precision

If additional external sensors are used to improve position estimation accuracy, then measurement precision is improved, but device complexity and cost increase

Engineering Contradiction:
Improveposition estimation accuracyVSAvoidsystem complexity
Core Design Contradiction:
Measurement precisionVSDevice complexity

Solution Approach 1:

The patent uses optical tracking systems during a preliminary training phase to create a trained AI system. Once trained, the system can operate using only the IMU, eliminating the need for continuous external sensors during actual use. This approach achieves high accuracy without permanently increasing device complexity.

Inventive Principle:
Principle #10Preliminary action

Solution Approach 2:

The trained AI system creates a computational model that copies the position and orientation determination capabilities of the optical tracking system. This virtual copy allows the system to achieve similar accuracy to external sensor-based systems without requiring those physical sensors during operation.

Inventive Principle:
Principle #26Copying

3Ease of manufacture

If inexpensive semiconductor sensors are used for IMU, then cost and size are reduced, but measurement precision deteriorates due to susceptibility to error and gravity interference

Engineering Contradiction:
Improvecost and sizeVSAvoidacceleration measurement accuracy
Core Design Contradiction:
Ease of manufactureVSMeasurement precision

Solution Approach 1:

The patent uses optical tracking data as feedback during the training phase to teach the AI system how to correct for sensor errors and gravity interference. The trained AI model learns to compensate for these issues in inexpensive semiconductor sensors, achieving high accuracy without requiring expensive high-precision sensors.

Inventive Principle:
Principle #23Feedback

Solution Approach 2:

The patent transforms the raw acceleration data from inexpensive sensors into accurate position and orientation information through the trained AI system. This parameter transformation process corrects for sensor imperfections and gravity interference, allowing low-cost sensors to achieve high measurement precision.

Inventive Principle:
Principle #35Parameter changes

Data Source

PatentEP4320409B1Method for calibrating and determining position data of an inertial measurement unit, training system, and medical instrument comprising an inertial measurement unit
Publication Date: 2025.12.10 B BRAUN NEW VENTURES GMBH
  • EP4320409B1 patent drawingFigure 1~3
  • EP4320409B1 patent drawingFigure 4~5b

AI summary

The present invention relates to a method for calibrating and predicting at least position data of an inertial measurement unit (2), which method is characterised by the following steps. using a motor-driven system (4) to move the IMU (2) along a predefined trajectory (T) in space; during the movement, using the IMU (2) to acquire measured movement data and providing the movement data to a control unit (8); during the movement, using a tracking system (6) to acquire a position and/or orientation of the IMU (2); linking/associating the measured movement data with the detected position and/or orientation in order to obtain a training data set; training an AI system (10) with the training data set in order to obtain an IMU calibration; detecting measured movement data of the IMU (2) as an input to the trained AI system (10); and, based on the input movement data, outputting a position and/or orientation of the IMU (2) using the trained AI system (10). The invention also relates to a training system (1), a medical instrument (18), a computer-readable storage medium, and a training data set according to the associated claims.