International Conference on Robotics and Mechatronics, ICROM 2013
Simultaneous Sensing and Actuating for Path Condition Monitoring of a Power Wheel Chair Hossein Mousavi Hondori
Pham Quoc Trung
Ling Shih-Fu
Department of Mechanical and Aerospace Engineering University of California, Irvine USA
[email protected]
School of Mechanical and Aerospace Engineering Nanyang Technological University Singapore
[email protected]
School of Mechanical and Aerospace Engineering Nanyang Technological University Singapore
[email protected]
Abstract—Monitoring the ambience’s condition is emerging as one of the most important issues in mobile robots. While we need to detect and process the condition quickly and efficiently using sensors, they (sensors) introduce extra complexity to the systems. Therefore this paper aims to develop a new method to simplify the sensing process. It describes an experimental process for monitoring path conditions of a power wheel chair inside and outside of the lab. The monitoring method triggers impedance via transduction matrix of the DC motors which run the wheelchair. The matrix is based on the relationship between electrical and mechanical impedance of the motors. A voltage probe and a current probe are used to monitor the changes of voltage and current supplied to the two DC motors of the system while the wheel chair is moving and the signals are recorded by a Digital Signal Analyzer (DSA). The recorded data are processed by MATLAB and the graphs representing various conditions are plotted. The relationship between the conditions of the path and the changes of voltage, current, and impedance of the motors are listed. The use of the method is justified but miniaturization of the voltage and current probes is necessary for real-life implementation of the system.
I. INTRODUCTION Impedance control is a concept proposed by Neville Hogan [1] which concerns robotics. It regards the importance of force control and in some way disregards the force trajectory. In other words it emphasizes on how the mechanical impedance is controlled and regulated. To implement impedance control, it is essential to use both velocity and force sensors. This adds up to the cost and intricacy of the systems. Anderson et al [2] used an approach to measure the mechanical impedance of a PZT coated cantilever beam derived from the electromechanical modeling of the actuator. The simultaneous sensing and actuating (SSA) was able to successfully measure and control the vibration of the PZT-coated beam. Ling et al [3], [4] were able to further advance the method by removing the loading effect caused by the shaker without solving the governing equations. Further more in [5–11] they applied the method to monitoring the load condition of a rotary motor. The SSA method has also been applied to moving electrical vehicles such as an omni-directional robot [12].
978-1-4673-5811-8/13/$31.00 ©2013 IEEE
The use of the method provided the capability of monitoring the conditions of the path the robot was undertaking. Similar to [12] this work will concern the path condition monitoring but it is also in continuation of our previous work [9] in the sense that it interests healthcare applications. Power wheel chairs are widely used by patients and people with disabilities. In most cases, inexpensive, fast, and reliable monitoring of the conditions of the wheelchair is desired. Monitoring the condition of the wheelchair can even inform the caregivers about patient’s daily routines of activities in a remoterehabilitation or home-rehabilitation setting. The motion condition monitoring process is as follows: 1) Measuring electrical impedance via measuring electrical voltage and current; 2) Obtaining the mechanical impedance of the path using transduction matrix; 3) Analyzing the data obtained from steps 1) and 2) in different path settings
II.METHODOLOGY A. Measuring Impedance In this paper electrical impedance and mechanical impedance are defined according to Eq. 1 and Eq. 2. The impedance is the measure of how the system resists against the flow of energy. For example a high impedance mechanical system will show lower velocity as compared to a low impedance mechanical system when both systems are subject to the same amount of force. The same concept (analogically) applies to electrical impedance. (1) E Ze = i (2) F Zm = v An electromechanical actuator can be modeled using a Transduction Matrix (TM) according to Fig. 1. As seen in Fig. 1, the matrix transforms the electrical signals to mechanical signals; it also shown in Eq. 3. Introducing Eq. 1 and Eq. 2 to the concept shown in Fig. 1 leads to the conclusion that the amount of mechanical impedance can be configured by measuring the electrical impedance. [7], [9].
International Conference on Robotics and Mechatronics, ICROM 2013
m7=2.0kg, d=50cm, t=8sec, E=9.4V, I=2.1A m8=1.0kg, d=49cm, t=8sec, E=9.4V, I=2.0A
Fig. 1. Transduction Matrix of the motor and mechanism
ìE ü éT11 T12 ù ìT ü í ý=ê ú´í ý îI þ ëT21 T22 û îw þ
(3)
B. Apparatus The system has two identical DC-motors at both sides of the wheelchair attached to two rear wheels and a joystick at front is used to adjust the direction as well as the speed of the wheel chair. The wheel chair will counter different speed levels and loading conditions inside and outside the lab such as turning and obstacles. When a wheel chair is moving along a path, there will be many different conditions such as slopes, obstacles etc. Whenever the path condition changes, the working parameters of the wheel chair will also change accordingly, these parameters include Torque (T) to drive the wheels and rotating speed (ω) of the wheels. The quotient of T/ω which is the mechanical impedance changes in different paths.
C. Obtaining the Transduction Matrix of the DC-Motor In SSA method, the relationship between Mechanical Impedance and Electrical Impedance depends on the characteristics of the DC-motor. Since each DC-motor has its own characteristics (or transduction matrix), the first work in this project is to find out the transduction matrix of the motors used on the wheel chair. In order to find out the transduction matrix, four signals will be measured including voltage (V), current (i), force (F), and velocity (v). To reduce the error, least square approximation method will be used. We measure 8 pairs of Voltage-Current value and 8 corresponding values of Force-velocity. To do so, 8 sets of different weights will be applied to the wheel and 8 corresponding sets of voltage and current will be recorded from the. During a fixed time t=8 sec, the linear distance in vertical direction is measured to calculate the velocity v. As seen in Fig. 2, in all the above tests, the wheelchair is places on top of a lab bench with the wheel taken out so that data acquisition becomes possible. These data are only used for obtaining the transduction matrix. Below are the recordings during the tests: m1=3.0kg, d=46cm, t=8sec, E=8.8V, I=2.2A m2=2.5kg, d=47cm, t=8sec, E=8.7V, I=2.1A m3=2.4kg, d=44cm, t=8sec, E=8.8V, I=2.1A m4=2.3kg, d=47cm, t=8sec, E=9.5V, I=2.1A m5=2.2kg, d=42cm, t=8sec, E=9.1V, I=2.1A m6=2.1kg, d=47cm, t=8sec, E=9.1V, I=2.1A
978-1-4673-5811-8/13/$31.00 ©2013 IEEE
Fig. 2. Experiment setup to find transduction matrix
In the data above m is the weight of the mass, d is the vertical distance the mass can move in the fixed time, t, during which the mass travels (t=8 sec). E is the mean value of the voltage measured at the terminals of the DC-motor and i is the current. From distance d and time t, we can find velocity v. Applying the above data into the Eq. 1 and Eq. 2, the electrical and mechanical impedance can be obtained. Moreover, introducing the data to Eq. 4, we can obtain the transduction matrix [Tij] using least squares approximation. é E1 L En ù éT11 T12 ù é T1 L Tn ù êI L I ú = ê ú´ê ú (4) nû ë 1 ëT21 T22 û ëw1 L wn û The transduction matrix is obtained by solving Eq. 4 in Matlab using least squares method; the result is seen in Eq. 5. éT11 T12 ù é0.1310 1.5138ù ê ú=ê ú (5) ëT21 T22 û ë0.1837 0.2916û Now that the transduction matrix is obtained, using the inverse of the transduction matrix, we can acquire torque and angular velocity values by measuring the voltage and current as shown in Eq. 6. -1 (5) ìT ü éT11 T12 ù ìE ü = ´ í ý ê í ý ú îw þ ëT21 T22 û îI þ
III. COLLECTION OF EXPERIMENTAL DATA A. Experimental Setup All measurement equipments including Digital Signal Analyzer (DSA), voltage probe and current probe are mounted on the seat of the wheel chair as shown in Fig. 3. The measurement equipments record the electrical signals in real time. The whole system will be allowed to move along the corridor, rotate and cross over an obstacle during the experiment.
International Conference on Robotics and Mechatronics, ICROM 2013
electrical power of the motor as the wheelchair passes through the obstacles. As shown in Fig. 5 when the wheel chair encounters the obstacles from the environment, there is a drop in electrical impedance and a jump in the power. These features of the signal may be used for path monitoring.
Fig. 3. Experimental setup in a corridor testing
B.
Experiment 1: Moving along a Straight Line
The wheel chair has 3 levels of speed: low, medium and high speed. At each speed level, 3 sets of data including Voltage and Current are recorded. The straight trajectory will cause the current and voltage of the motor to be nearly constant hence to save space we do not present the plots and only present the mean values for voltage and current as shown in Table 1 and Table 2. Table 1. The electrical data obtained during the experiments Speed Voltage Current Power Electrical (v) (A) (watt) impedance(Ω) Low 9.5 2.1 19.95 4.52 Medium 15 2.25 33.75 6.66 High 20.5 2.45 50.225 8.36 Table 2. The mechanical data obtained during the experiments Torque Angular Mechanical Speed velocity impedance (N.m) (rad/s) (N.m.s) Low 1.7041 6.1281 0.2781 Medium 4.0351 10.2580 0.3934 High 9.4586 14.3606 0.6586 Comparing the medium to the low speed: The voltage increases 57.8% (high percentage), the current increases 7.2% (low percentage compared to changing rate of voltage). The electrical power increases 69.2%. The electrical impedance increases 47.3%. Comparing the high to the medium speed: The Voltage increases 36.6% (high percentage), the Current increases 8.8% (low percentage compared to the changing rate of Voltage). The Electrical Power increases 48.8%. The Electrical Impedance increases 25.5%
C. Crossing Obstacles In this experiment, the wheelchair passes from some obstacles on the ground. The obstacles are pieces of flat piece of plastic (size 400x200x10 mm). Fig. 4 and Fig. 5 show the voltage, current, electrical impedance, and
978-1-4673-5811-8/13/$31.00 ©2013 IEEE
Fig. 4. Voltage and current when crossing obstacle 1
Fig. 5. Power and impedance when crossing obstacle 1
Now the Mechanical Impedance (or the obstacle) is presented in the Electrical Impedance form. Here the Power increases from 20W to 25W (25% increased), and the Electrical Impedance decreases from 9Ω to 6Ω (33% decreased). Another two random obstacle conditions were created and the measurement results led us to the similar features that observed in Fig. 4 and Fig. 5.
IV.DISCUSSION AND CONCLUSION From all the experiments done in different speed conditions and with several obstacles, we observed that encountering obstacles will mainly affect current of the system which causes both power and impedance to changes. The sharp changes in power and impedance signals can be considered as features that can be easily detected by the system to monitor the path conditions. Capabilities of the system can be use in the following applications: -Monitoring the impedance and power in different terrains such as asphalt, sand, mud etc. -Monitoring the impedance and power in a home setting while the patients is using the wheel-chair for moving around the home. Recording of such data and analyze them throughout a day will enable doctors to monitor the general well-being of the patient especially if
International Conference on Robotics and Mechatronics, ICROM 2013
the elderly lives in a home alone which is common-place case in the USA and Singapore where this study is conducted. For example the doctor can monitor how much the patient has moved around the day; not moving for several hours can be an indication of an urgent matter so they can call the patient to check on his/her health. -These data can help controlling the wheelchair in slopes where electrical power consumed suddenly drops. When the patient drives his/her wheelchair to a down-hill slope there is high risk of losing control will can cause serious damage. -Another application is that if these data are monitored properly so that we can distinguish between a functioning wheelchair and wheelchair that is soon going to need maintenance, the care givers or hospital can prevent the system from being damaged by timely maintenance. For future work, miniaturization of the voltage and current sensors is the most important thing to do. Also using small A/D signal acquisition devices that connect to cell phones through Bluetooth can transmit the data to the mobile phone nearby which will in turn send the data to the hospital. Using smart phone apps (i.e. Andoird or iPhone) can even give feedback on the patient directly. The comprehensive goal of this work is to be part of a wider scope of affordable home-based robotic and nonrobotic rehabilitation [13–15] setting for monitoring [16] and training [17], [18] of patients with motor disabilities.
[7]
V.REFERENCES
[14]
[1] [2]
[3]
[4]
[5]
[6]
N. Hogan, “Impedance Control: An Approach to Manipulation,” in American Control Conference, 1984, 1984, pp. 304 –313. E. H. Anderson and N. W. Hagood, “Simultaneous Piezoelectric Sensing/Actuation: Analysis And Application To Controlled Structures,” Journal of Sound and Vibration, vol. 174, no. 5, pp. 617–639, Jul. 1994. S.-F. Ling and Y. Xie, “Detecting mechanical impedance of structures using the sensing capability of a piezoceramic inertial actuator,” Sensors and Actuators A: Physical, vol. 93, no. 3, pp. 243–249, Oct. 2001. S.-F. Ling, X. Hou, and Y. Xie, “Decoupling loading effect in simultaneous sensing and actuating for dynamic measurement,” Sensors and Actuators A: Physical, vol. 120, no. 1, pp. 257–265, Apr. 2005. L. Fu, S.-F. Ling, and C.-H. Tseng, “On-line breakage monitoring of small drills with input impedance of driving motor,” Mechanical Systems and Signal Processing, vol. 21, no. 1, pp. 457–465, Jan. 2007. H. M. Hondori and S. F. Ling, “A Novel Device for Measuring Mechanical Impedance during Dynamic Tasks,” presented at the BIODEVICES 2010, 2010, pp. 64–68.
978-1-4673-5811-8/13/$31.00 ©2013 IEEE
[8]
[9]
[10] [11]
[12]
[13]
[15]
[16]
[17]
[18]
Hossein Mousavi Hondori, “Measuring Human Arm’s Mechanical Impedance for Assessment of Motor Function,” Nanyang Technological University, Singapore, 2012. H. M. Hondori and L. Shih-Fu, “Perturbation-based measurement of real and imaginary parts of human arm’s mechanical impedance,” in 2010 Annual International Conference of the IEEE Engineering in Medicine and Biology Society (EMBC), 2010, pp. 5911 –5914. H. M. Hondori and L. Shih-Fu, “A method for measuring human arm’s mechanical impedance for assessment of motor rehabilitation,” in Proceedings of the 3rd International Convention on Rehabilitation Engineering & Assistive Technology, New York, NY, USA, 2009, pp. 17:1–17:4. H. M. Hondori and L. S. Fu, “A Simultaneous Sensing cum Actuating Method for Measuring Human Arm’s Mechanical Impedance.” H. M. Hondori and L. Shih-Fu, “Monitoring Human Arm’s Mechanical Impedance while Adapting a Reaching Task.” Y. L. Kuo, J. Y. Lin, and Y. N. Lee, “Motion conditions justification of an omni-directional mobile platform by electro-mechanical transduction matrix,” Transactions of the Institute of Measurement and Control, vol. 31, no. 2, pp. 167–179, Apr. 2009. H. M. Hondori, M. Khademi, L. Dodakian, S. C. Cramer, and C. V. Lopes, “A Spatial Augmented Reality Rehab System for Post-Stroke Hand Rehabilitation,” presented at the 2013 Conference on Medicine Meets Virtual Reality, NextMed/MMVR20, 2013. Hossein Mousavi Hondori, Maryam Khademi, and Cristina Videira Lopes, “Use of a Portable Device for Measuring Arm’s Planar Mechanical Impedance during Motion,” in 2012 IEEE EMBS Conference on Biomedical Engineering and Sciences (IECBES). H. M. Hondori and A. W. Tech, “Smart mug to measure hand’s geometrical mechanical impedance,” in 2011 Annual International Conference of the IEEE Engineering in Medicine and Biology Society,EMBC, 2011, pp. 4053 – 4056. H. M. Hondori, M. Khademi, and C. V. Lopes, “Monitoring Intake Gestures using Sensor Fusion (Microsoft Kinect and Inertial Sensors) for Smart Home Tele-Rehab Setting,” in 2012 1st Annual IEEE Healthcare Innovation Conference, Houston, TX, 2012. M. Khademi, H. M. Hondori, Cristina Videira Lopes, Lucy Dodakian, and Steve C. Cramer, “Haptic Augmented Reality to Monitor Human Arm’s Stiffness in Rehabilitation,” in 2012 IEEE EMBS Conference on Biomedical Engineering and Sciences (IECBES), 2012. H. M. Hondori, A. W. Tech, M. Khademi, and C. V. Lopes, “Real-time Measurement of Arm’s Mechanical Impedance with Augmented Reality Illustration,” 2012.
