Osama Majdalawieh

A new efficient control algorithm using potential field: extension to robot path tracking

Intensive study has been given to robot navigation in the past decade. Among various proposed contributions in robot study, potential field (PF) methods enjoy a great reputation in robotics applications due to their simplicity, elegance, and high efficiency. However, despite the intensive study in designing suitable PFs for real applications and raising possible remedies for their inherent limitations with the aid of other means, few works have concentrated on accurate tracking motion control using PFs. The problem of controlling wheeled mobile robot platforms to track a path is studied. We take our inspiration from the concept of potential fields, give a different interpretation to them, and then propose a new efficient tracking control algorithm using VOPF (virtual obstacle potential field) since obstacles nearby the robot are imaginary. Meanwhile, this method takes advantage of the strengths of PFs and the virtual vehicle approach. Based on mathematical analysis, the paper presents its feasibility. Also, the methodology can be applied to any mobile robot since it is model independent.

Biomedical signal processing and rehabilitation engineering: a review

Combining biomedical signal processing and rehabilitation engineering can yield sophisticated and technological designs that will extend the capabilities and improve the quality of life of the disabled person. Information obtained from a biomedical signal, such as electroencephalogram (EEC), can be used to select the appropriate assistive device for the disabled person. In this paper, a thorough review and discussion of state of the art technology in biomedical signal processing and rehabilitation engineering is presented. Various signal processing methods and techniques have been applied to biomedical signals. Towards the end of the paper, challenges and future trends are discussed.

Output vibration measurements of bone-anchored hearing AIDS.

Hypothesis: Different bone-anchored hearing aids (BAHAs) processors have different output vibration characteristics, which depend on the mechanical load and the volume setting. Responses will differ between live heads and dry or plastic skulls. Background: The BAHA is an implantable bone-conduction device. Three different BAHA models are available. Their output vibrations have not been reported using a noncontact method with differing impedance loads, including the BAHA-fitted patient head. Methods: Using a laser-Doppler vibrometer, vibration responses with sound input of 70- to 80-dB sound pressure level were measured on unloaded BAHAs, a dry skull, a plastic skull, and on the abutments of three live BAHA-fitted patients. Responses at different volume settings and distances from the vibrator were also tested. Frequency responses were calculated for displacement, velocity, and acceleration. Results: Unloaded BAHA accelerations were approximately 30 to 50 dB higher than live-head accelerations. Live-head accelerations were similar to dry skulls in frequencies of more than 500 Hz, but much higher than the plastic skull responses. Live-head responses were more damped. The Cordelle II outperformed the other two processors by approximately 20 dB. The Classic 300 had better low-frequency responses than the Compact. The volume settings had little effect on vibration output overall. Acceleration peak was at approximately 2.5 kHz for all conditions. Conclusion: The BAHA processors differ in the output acceleration they can achieve with differing loads. The volume control setting has little impact on accelerations produced for most processors. The live-head responses are similar to the dry skull in frequencies of more than 500 Hz.

An HTK-developed hidden Markov model (HMM) for a voice-controlled robotic system

This paper presents a voice-controlled mobile robotic system capable of recognizing voice commands and relaying them to a mobile robot. First, the specifications of the robot used will be presented. Second, a description of the HTK toolkit, the HMM, and the VCR software developed will be discussed. Finally, comparing our HTK-developed HMM against a commercially available Microsoft speech recognition engine (SDK 5.1) in terms of accuracy. The experimental evaluation and accuracy tests will show the ability of the VCR software to control a robot with simple human voice commands. Conclusions and future work are presented towards the end of the paper.

The effect of increased stiffness of the incudostapedial joint on the transmission of air-conducted sound by the human middle ear

Objective To study the possible effects of increased stiffness at the incudostapedial joint (ISJ) on sound transmission in the human middle ear. Background The physiologic role played by the IS joint in the mechanics of human middle ear function is unclear. It is also unclear how fixation of this joint might manifest itself and what the implications are of fixing this joint during surgical reconstruction. Hypothesis Increased stiffness of the ISJ will affect sound transmission through the middle ear. Methods Cyanoacrylate adhesive was instilled around the ISJ joint in 5 fresh human cadaveric temporal bones to increase ISJ stiffness. Laser Doppler vibrometry measurements of sound-induced peak-to-peak displacement of the umbo and stapes footplate were made before and after stiffening the ISJ. Results At baseline, the measurements at the stapes footplate followed those at the umbo but were approximately 12 dB lower in the speech frequencies. We found that stiffening the ISJ produced an almost equal decrease in peak-to-peak displacement at both the umbo and the stapes footplate, with little change in their relative motion, consistent with an increase in ossicular impedance. The decrease was mainly between 400 and 1,000 Hz with a statistically significant mean magnitude loss of 6 dB at 740 Hz. Conclusion Increased stiffness at the ISJ results in a small, probably clinically insignificant decrease of 3 to 6 dB in middle ear sound transmission in the lower frequencies between 400 and 1,000 Hz.

Linearity and Lever Ratio of the Normal and Reconstructed Cadaveric Human Middle Ear

Hypothesis The linearity and the level ratio are different in reconstructed ears. Background The linearity of the reconstructed human middle ear (ME) has not been previously explored. It is important to analyze if high sound pressure levels (SPLs) result in distortion due to nonlinearities particularly because hearing aids have high-output SPLs. The diseased ME is reconstructed with prostheses. These diseased ears frequently need additional amplification with hearing aids, and it is unclear if reconstruction itself leads to nonlinear ME responses. Methods Eight fresh human cadaveric temporal bones were used. Pure tones of 70, 90, and 110 dB SPL at 500, 1,000, and 3,000 Hz were presented to the ear canal. Umbo and stapes displacements were measured by means of a laser Doppler vibrometer. After removing the incus, the tympanic membrane assembly to the stapes head prosthesis was placed, and measurements were repeated. Results Stapes footplate vibrations in the reconstructed ears are 10 to 15 dB lower than those of the normal ears. In both normal and reconstructed ears, the footplate vibrations are linearly related to SPL at the tympanic membrane between 70 and 110 dB SPL at the frequencies tested. For the lever ratio, intact ears are more efficient at transmission of umbo vibrations to the stapes compared with reconstructed ears. Conclusion To within acceptable limits, the ME seems to be linear between 70 and 110 dB SPL input levels, across the speech frequencies, and this does not change with reconstruction. The reconstructed human ME seems to have a less efficient lever ratio than the intact ME.

Is there loss of vibration amplitude across the snap coupling of the bone-anchored hearing aid?

Hypothesis: There is loss of vibration transmission across the snap coupling connecting the Bone-Anchored Hearing Aid transducer to the implanted abutment on the head. Background: The only nonrigid part of the Bone-Anchored Hearing Aid system is the connection between the output of the transducer and the abutment. Vibration losses across the coupling have not been previously measured. If a loss is found, a change in design could improve the efficiency of the Bone-Anchored Hearing Aid. This would be very helpful in borderline cases in which the Bone-Anchored Hearing Aid does not have enough power to achieve adequate hearing threshold levels. Methods: A laser Doppler vibrometer was used to measure vibrations on the output stem and four points on the abutment of the Bone-Anchored Hearing Aid. The Bone-Anchored Hearing Aid was coupled to a dry skull through a plexiglas bite bar screwed to the skull. The impedance load was varied by fixing the skull. A control loose coupling was measured. Five Bone-Anchored Hearing Aid Compacts were measured. Results: There was little loss across the Bone-Anchored Hearing Aid snap coupling. At frequencies above 500 Hz, there was no more than 5-dB loss at any frequency. Changing the impedance load by fixing the skull did not change the loss across the coupling. Conclusion: The snap coupling is an efficient means of transmitting vibrations to the skull. There is little loss of vibration attenuation across it. Increases in functional Bone-Anchored Hearing Aid amplification gain cannot be achieved by further optimizing this interface.

A new landmark framework for mobile robot localization and asynchronous sensor fusion

This paper presents a new landmark framework for mobile robot localization in which we use a combination of obstacle vertical edge and its corresponding plane angle, that appear in practically all indoor environments, to realize mobile robot localization. Sensor data is assumed to be provided in the form of a Gaussian distribution over the space of the robot poses. We propose a new fusion strategy asynchronous sensor fusion (ASF), which either creates complementaries or adds more redundancy for asynchronous sensor data set will also be presented in this paper. This strategy, combined with dead reckoning, enables mobile robots to asynchronously utilize incomplete information captured by range sensors to localize themselves and simultaneously accomplish map building in a partially known indoor environment.