Use of inertial sensors in lower-limb rehabilitation device.

Dumitriu, Adrian ; Barbu, Daniela Mariana

1. INTRODUCTION

A novel lower-limb rehabilitation device equipped with suitable inertial sensors is presented, which enables training of a leg affected by strokes or injuries, in coordination with the movements of the other, normal, leg. This simple 1-DOF device can be included in the category of therapy robots.

As shown in (Van der Loss & Reinkensmeyer, 2008), therapy robots have at least two main users simultaneously, the person with a disability who is receiving the therapy and the therapist who sets up and monitors the interaction with the robot. An important number of these devices are destined for upper-and lower-limbs movement therapy. A robot may be a good alternative to a physical or occupational therapist for the actual hands-on intervention for several reasons: (1) once properly set up, an automated exercise machine can consistently apply therapy over long periods of time without tiring; (2) the robot's sensors can measure the work performed by the patient and quantify, to an extent perhaps not yet measurable by clinical scales, any recovery of function that may have occurred, which may be highly motivating for a person to continue with the therapy; and (3) the robot may be able to engage the patient in types of therapy exercises that a therapist cannot do.

Many attemps are made to develop new motion tracking systems to support the rehabilitation programme for pacients at home, so that the burden and hospitals can be relieved, and inertial sensors seem to offer proper solutions for this goal.

(Tao et al., 2007) introduce a real-time hybrid solution to articulate 3D arm motion tracking for home based rehabilitation by combining visual and inertial sensors. For inertial motion tracking a MT9 inertial sensor from Xsens has been used.

(Zhou et al., 2008) developed a home based rehabilitation system equipped with body-mounted inertial sensors, able to provide measurements of upper limb motion. To estimate the position of the shoulder joint, a Lagrangian based optimization technique has been used, which integrates the values of acceleration and the estimated value of rotation measured by two inertial sensor units. Each unit consists of a tri-axial accelerometer, a tri-axial gyroscope and a tri-axial magnetometer.

(Azevedo & Heliot, 2005) use inertial micro-sensors which associate 3 accelerometers and 3 magnetometers in a minimal volume, to identify postural task transition, as soon as possible after the patient makes a decision, in order to allow for optimal posture preparation and execution and compare the execution of the ongoing task with the reference pattern, in order to identify the current movement phase.

[FIGURE 1 OMITTED]

2. REHABILITATION DEVICE STRUCTURE

2.1 Role and configuration

The rehabilitation device proposed is a simple one-degree-of-freedom (1-DOF) mechanism attached, with proper splints and straps, to the patient's leg, which must be trained due to movement and coordination problems, as effect of strokes or injuries. For simplicity, the device is considered as a rigid link pivoting about the knee joint and powered by a single actuator. Actuators tested are D.C. geared motors with brushes, at 12 V, with a nominal torque of 1,2 Nm and an output speed of either 20 rpm or 100 rpm. For the motor with 100 rpm an external gear between motor shaft and mobile element of the device amplifies the motor torque to 6 Nm at 20 rpm. Adjustable switches mounted on the joint serve to limit rotation angles of the trained leg according to the therapy strategy and to avoid injuries due to machine failures.

The block diagram of the device is shown in Fig.1. Sensors attached to the active leg measure accelerations, angular velocities and rotation angles of movements and the control computer commands the DC motor to assure a similar movement of the trained leg, based on the feed-back from sensors attached to this leg.

2.2 Sensor units used for prototype

Study of kinematics and dynamics of the two legs, the active and the trained, lead to following conclusions:

* Both legs must be equipped with sensors able to measure movement angles in knee joints;

* Angular velocity in sagittal plane of active and trained leg must be measured, in order to assure movements with same angular velocities in both legs;

* Measure of accelerations in sagittal and frontal planes of both legs are useful for an efficient control.

Due to the fact that two MTx miniature 3DOF inertial sensor units from Xsens Motion Technologies were available in the laboratory, they have been used for experiments with the prototype device. These sensor units provide drift-free 3D orientation as well as kinematic data: 3D acceleration (MEMS solid state sensors, capacitive readout), 3D rate gyro (MEMS solid state sensors, monolithic, beam structure, capacitive readout) and 3D earth-magnetic field (magnetometer in thin film magneto resistive technology) and are excellent measurement units for orientation and kinematic data measurements of human body segments. One MTx unit has been attached using suitable straps to the shank of each leg, one mounted directly on the active leg and the other fixed on the mobile metallic element of the rehabilitation device.

[FIGURE 2 OMITTED]

Fig.2 presents the acquisition and control scheme of the prototype device, organized around a notebook PC: 2 USB interfaces connect the MTx sensor units, via USB converters and a USB/RS-232 converter allows communication with the DC motor control board. C++ software has been used to develop the acquisition and control program.

3. SOLUTION WITH DISTRIBUTED SENSORS

MTx inertial measurement units were very useful in the development phase, because they allowed measurement of kinematic parameters in 3D space and comparison of knee angles directly measured and those computed using two other methods:

* Integration of angular velocity in the sagittal plane;

* Calculus of inclination using acceleration measured in frontal plane.

These sensor units are too complex and expensive for usual rehabilitation devices and, based on the experience gained, solutions have been developed and tested, based on simple and cheep acceleration and gyroscope sensors in MEMS technology.

3.1 Acceleration sensors

Memsic 2125 acceleration sensor from Parallax Inc. has been used, because its simplicity, both from hardware and software point of view. This sensor, in MEMS technology, contains, internally, a small heater. This heater warms a "bubble" of air within the device. When gravitational forces act on the bubble, it moves and this movement is detected by very sensitive thermopiles (temperature sensors). On-board electronics convert the bubble position (relative to g-forces) into pulse outputs for the X and Y axes. Main characteristics:

* Measure 0 to [+ or -] 2g on either axis, with less than 1 mg resolution;

* Simple, pulse (PWM) outputs of g-force for X and Y axis, requiring just two I/O pins.

A lot of experiments for determing inclination angles with this type of sensors have been carried out by the author for mobile robots navigation (Dumitriu, 2008).

Two sensors have been mounted adequately on the tigh and shank of each leg and connected to 4 available I/O ports of a microcontroller board, one for each pulse output of g-force for x and y axis of a sensor.

3.2 Gyro sensors

The basic operating element of a MEMS rate gyroscope is a flexing piezoelelectric beam coupled to a sensing element. The input voltage to the sensor is used to drive the piezoelectric element to resonance. Rotation about the axis of the beam induces a Coriolis force, proportional to the mass and angular velocity of the sensing element, which is detected by a capacitive element which seizes the Coriolis displacement of the vibrating beam. MEMS gyroscopes with integrated signal processing electronics in a single silicon piece are now widely available, but much of them require higher input voltages (12-16 V) than those supported by standard acquisition systems. The ADXRS300 from Analog Devices has been used, because this chip includes circuitry with a charge pump to amplify the input voltage from 5 V to the required 12-16 V. ADRS300 is a [+ or -] 300 single axis rate gyro with signal conditioning. The output of a differential displacement capacitor is sent through demodulation stages included in circuitry on the chip, resulting in an analog voltage output, proportional to angular rate.

An ADXRS300 is mounted on the shank of each leg with the active (yaw) axis in sagittal plane.

Knee angles are computed by integration of angular velocities measured with gyro sensors. For safety they are compared with angles calculated using data from acceleration sensors.

4. CONCLUSIONS AND FUTURE WORK

Experiments have proved that gyro and acceleration sensors are suitable for motion tracking and for controlling a lower-limb rehabilitation device.

Future work will be focused on two objectives:

* Use of a data base with memorized kinematic parameters for optimizing training with the rehabilitation device;

* Use of sensors (current and acceleration sensors) for measuring the work performed by the patient and quantifying recovery of trained leg during the therapy.

5. ACKNOWLEDGEMENTS

This is a part of the project "Contributions to analysis, modelling and simulation of modern mechatronic systems destined to medical recovery", founded by Romanian Ministry of Education, Research and Innovation, under Grant ID_147/2009.

6. REFERENCES

Azevedo, Ch. & Heliot, R. (2005). Rehabilitation of Functional Posture and Walking: Coordination of Healthy and Impaired Limbs, Available from: http://intechweb.org, Accessed: 2009-04-15

Dumitriu, A. (2008). Using Fuzzy Reasoning to Guide a Mobile Robot on a Rough Terrain, Annals of DAAAM for 2008 & Proceedings of the 19th International DAAAM Symposium, Katalinic, B. (Ed.), pp. 230-231, ISBN 978-3-901509-68-1, Trnava, Slovacia, October 2008, Published by DAAAM International, Vienna

Tao, Y.; Hu, H. & Zhou, H. (2007). Integration of Vision and Inertial Sensors for 3D Arm Motion Tracking in Home-Based Rehabilitation. The International Journal of Robotics Research, Vol.26, No.6, (June 2007), pp.607-624, ISSN: 0278-3649

Van der Loss, H. F. M. & Reinkensmeyer, D. J. (2008). Rehabilitation and Health Care Robotics (Chapter 53). In: Springer Handbook of Robotics, Siciliano & Khatib, (Ed.), pp.1223-1251, Springer Verlag, ISBN 978-3-540-23957-4, Berlin, Heidelberg

Zhou, H.; Stone, T.; Hu, H. & Harris, N. (2008). Use of Multiple Wearable Inertial Sensors in Upper limb Motion Tracker. Medical Engineering & Physics, Vol.30, (2008) pp.123-133, ISSN: 1350-4533

*** (2009) http://www.analog.com--[+ or -] 300[degrees]/s Single Chip Yaw Rate Gyro with Signal conditioning, Accesed on:2009-04 10

*** (2008) MTi and MTx User Manual and Technical Documentation, Xsens Motion Technologies, Netherlands, Document MT0100P, Revision J, April 1, 2008