System for capturing movements of an articulated structure

• US 9,958,470 B2
• Filed: 03/07/2013
• Issued: 05/01/2018
• Est. Priority Date: 03/08/2012
• Status: Active Grant

• Current Assignee: Commissariat a L'Energie Atomique (AES Corporation), Movea
• Original Assignee: Commissariat a L'Energie Atomique (AES Corporation), Movea
• Inventors: Grenet, Pierre
• Primary Examiner(s): Raymond, Edward
• Application Number:
  US14/382,998
• Publication Number:
  US 20150032408A1
• Time in Patent Office: 1,881 Days
• Field of Search: 702141, 702150, 702182-185
• CPC Class Codes: A61B 5/1116 Determining posture transitions
• CPC Class Codes: A63B 71/06 Indicating or scoring devices for games or players , or for other sports activities
• CPC Class Codes: G01P 15/00 Measuring acceleration; Measuring deceleration; Measuring shock, i.e. sudden change of acceleration
• CPC Class Codes: G06F 3/011 Arrangements for interaction with the human body, e.g. for user immersion in virtual reality blind teaching G09B21/00

Abstract

A system for capturing the movements of a body having substantially rigid segments articulated together includes attitude units fastened onto the segments of the body, the units each including at least one accelerometer and one magnetometer, and a reduced number of gyroscopes. The system also includes a pseudo-static state detection module and a module for calculating pseudo-static angles. When all segments are detected in a pseudo-static state, the state vector is provided by the module for calculating pseudo-static angles. When a segment is detected in a dynamic state, the state vector is provided at the output of a Kalman filter.

Claims

1. A system for capturing movements of a structure including at least N substantially rigid segments articulately connected with said structure, said system including:

a set of N accelerometers with at least one measurement axis, each of said N accelerometers being substantially rigidly connected to one of said N segments,

a set of P second sensors capable of returning one direction of a fixed reference frame, each substantially rigidly connected to a segment,

a set of Q third sensors capable of returning a measurement representative of an angular velocity, each substantially rigidly connected to a segment,

a module for communicating the outputs of the N accelerometers, P second sensors and Q third sensors with a computer processing module;

said processing module including a state observer,

said system further including;

a pseudo-static state detection module for detecting a pseudo-static state of each of the segments of said structure,

a pseudo-static orientation calculation module for calculating a pseudo-static orientation of the segments in a pseudo-static state,

a state observer module configured for replacing outputs of a prediction function of the state observer with outputs of the pseudo-static orientation calculation module for the segments for which a detection condition at the output of the pseudo-static state detection module condition is true

wherein the number Q is less than the number N and the number P.

2. The system for capturing movements as claimed in claim 1, wherein R is equal to a number of branches of the structure whereof the movement is captured, the number Q being less than or equal to R+1.

3. The system for capturing movements as claimed in claim 1, wherein the second sensors are magnetometers.

4. The system for capturing movements as claimed in claim 3, wherein the third sensors are gyroscopes.

5. The system for capturing movements as claimed in claim 4, wherein the observer is a Kalman filter.

6. The system for capturing movements as claimed in claim 5, wherein the modules thereof are configured by state evolution models of the form:

7. The system for capturing movements as claimed in claim 6, wherein the state vector x is of the form x=[θ

, {dot over (θ

)}, {umlaut over (θ

)}, Acc_x, Acc_y, Acc_z] where θ

, {dot over (θ

)} and {umlaut over (θ

)} respectively designate an orientation angle of the segments, its first derivative and its second derivative with respect to time, and where Acc_x Acc_y and Acc_z designate components of a frame acceleration Acc of the whole of the body in a terrestrial reference frame (X, Y, Z).

8. The system for capturing movements as claimed in claim 6, wherein the state evolution model of the Kalman filter uses said pseudo-static angles for the segments detected in a pseudo-static state and the state evolution model of the Kalman filter for the segments detected in a dynamic state.

9. The system for capturing movements as claimed in claim 8, wherein the state evolution model of the Kalman filter uses an assumption of constancy of accelerations of the articulation angles.

10. The system for capturing movements as claimed in claim 6, wherein the state vector is estimated from pseudo-static angles at the output of the pseudo-static orientation calculation module, if all the segments are detected in a pseudo-static state and is estimated by the Kalman filter, if at least one segment is detected in a dynamic state.

11. The system for capturing movements as claimed in claim 6, wherein a criterion of pseudo-staticity is fulfilled by a segment when at least one of the values provided by at least one of the elements of the group including an attitude unit or a gyroscope which is rigidly connected thereto provides at least one measurement chosen from the norm of an acceleration vector and an angle between said acceleration vector and a magnetic field vector which is less than a predetermined threshold value.

12. The system for capturing movements as claimed in claim 6, wherein the state evolution model of the Kalman filter predicts the angle θ

, its first derivative {dot over (θ

)} and its second derivative {umlaut over (θ

)} via the function defined by;

13. The system for capturing movements as claimed in claim 6, wherein the state evolution model of the Kalman filter predicts the frame acceleration Acc as a mean of frame accelerations Acc_overall calculated for each accelerometer from a measurement vector Mes_Acc_actual actually provided by said accelerometer, the frame acceleration Acc_overall being calculated for each accelerometer by:

Acc_overall=−

^EarthR_Sensor*(Mes_Acc_actual−

Mes_Acc_{Acc_overall=0})

Where ^EarthR_Sensor designates a matrix of rotation of a reference frame associated with said accelerometer toward the terrestrial reference frame, and Mes_ACC_{Acc_overall=0} designates a theoretical measurement vector calculated by assuming that the overall acceleration is zero.

14. The system for capturing movements as claimed in claim 6, wherein the accelerometer measurement model of the Kalman filter is given by:

15. The system for capturing movements as claimed in claim 6, wherein covariances of the measurement and state evolution noise are estimated both a priori and a posteriori.

16. The system for capturing movements as claimed in claim 6, wherein covariances of state evolution noise are estimated a priori from a comparison between predicted states and actual states taken from a database of body movements.

17. The system for capturing movements as claimed in claim 6, wherein covariances of state evolution noise are re-estimated a posteriori by calculating for each segment a gain index for the state evolution model from differences between standard deviations of the sensors for actual states taken from a database of body movements and deviations measured by said sensors.

18. A method for capturing movements of a structure including at least N substantially rigid segments articulately connected with said structure, said method including:

a step of acquiring measurements from N accelerometers with at least one measurement axis, said accelerometers being substantially rigidly connected to said segments,

a step of acquiring measurements from P second sensors configured to return a direction of a fixed reference frame, each substantially rigidly connected to a segment,

a step of acquiring measurements from Q third sensors configured to return a measurement representative of an angular velocity, each substantially rigidly connected to a segment, and

a step of communicating outputs of the N accelerometers, the P second sensors and the Q third sensors to a computer processing step comprising;

a step of detecting a pseudo-static state of each of the segments of said structure,

a step of calculating a pseudo-static orientation of the segments in a pseudo-static state, and

if at least one output of the step of detecting a pseudo-static condition is false, a step of replacing outputs of a prediction function of a state observer which receives as input the outputs of the N accelerometers, the P second sensors and the Q third sensors, Q being less than the number N and the number P, with the outputs of the step of calculating the pseudo-static orientation for the segments for which the detection condition at an output of the pseudo-static state detection module is true.