SYSTEM FOR CAPTURING MOVEMENTS OF AN ARTICULATED STRUCTURE
First Claim
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 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 truewherein the number Q is less than the number N and the number P.
2 Assignments
0 Petitions
Accused Products
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.
-
Citations
18 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 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. - View Dependent Claims (2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, 16, 17)
-
-
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
)}, Accx, Accy, Accz] 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 Accx Accy and Accz designate components of a frame acceleration Acc of the whole of the body in a terrestrial reference frame (X, Y, Z).
- , {dot over (θ
-
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;
- , its first derivative {dot over (θ
-
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=−
EarthRSensor*(Mes_Acc_actual−
Mes_AccAcc— overall=0)Where Earth RSensor designates a matrix of rotation of a reference frame associated with said accelerometer toward the terrestrial reference frame, and Mes_ACCAcc — 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.
-
Specification