ABSTRACT
In the automotive field, the introduction of keyless access systems is revolutionizing car entry techniques currently dominated by a physical key. In this context, this paper investigates the possible use of smartphones to create a PEPS (Passive Entry Passive Start) system using the BLE (Bluetooth Low-Energy) Fingerprinting technique that allows, along with a connection to a low-cost BLE micro-controllers network, determining the driver's position, either inside or outside the vehicle. Several issues have been taken into account to assure the reliability of the proposal; in particular, (i) spatial orientation of each microcontroller-based BLE node which ensures the best performance at 180° and 90° referred to as the BLE scanner and the advertiser, respectively; (ii) data filtering techniques based on Kalman Filter; and (iii) definition of new network topology, resulting from the merger of two standard network topologies. Particular attention has been paid to the selection of the appropriate measurement method capable of assuring the most reliable positioning results by means of the adoption of only six embedded BLE devices. This way, the global accuracy of the system reaches 98.5%, while minimum and maximum accuracy values relative to the individual zones equal, respectively, to 97.3% and 99.4% have been observed, thus confirming the capability of the proposed method of recognizing whether the driver is inside or outside the vehicle.
ABSTRACT
The paper deals with the problem of improving the maximum sample rate of analog-to-digital converters (ADCs) included in low cost wireless sensing nodes. To this aim, the authors propose an efficient acquisition strategy based on the combined use of high-resolution time-basis and compressive sampling. In particular, the high-resolution time-basis is adopted to provide a proper sequence of random sampling instants, and a suitable software procedure, based on compressive sampling approach, is exploited to reconstruct the signal of interest from the acquired samples. Thanks to the proposed strategy, the effective sample rate of the reconstructed signal can be as high as the frequency of the considered time-basis, thus significantly improving the inherent ADC sample rate. Several tests are carried out in simulated and real conditions to assess the performance of the proposed acquisition strategy in terms of reconstruction error. In particular, the results obtained in experimental tests with ADC included in actual 8- and 32-bits microcontrollers highlight the possibility of achieving effective sample rate up to 50 times higher than that of the original ADC sample rate.
ABSTRACT
The study reported here explored the design and realization of a low-cost, electromyographically controlled hand prosthesis for amputees living in developing countries. The developed prosthesis is composed of a light aluminum structure with opposing fingers connected to a DC motor that imparts only the movement of grasp. Problems associated with surface electromyographic signal acquisition and processing, motor control, and evaluation of grasp force were addressed, with the goal of minimizing cost and ensuring easy assembly. Simple analog front ends amplify and condition the electromyographic signals registered from two antagonist muscles by surface electrodes. Analog signals are sampled at 1 kHz and processed by a microcontroller that drives the motor with a supply voltage proportional to the muscular contraction, performing the opening and closing of the opposing fingers. Reliable measurements of the level of muscle contractions were obtained by specific digital processing: real-time operators implementing the root mean square value, mean absolute value, standard deviation, and mean absolute differential value were compared in terms of efficiency to estimate the EMG envelope, computational load, and time delay. The mean absolute value operator was adopted at a time window of 64 milliseconds. A suitable calibration procedure was proposed to overcome problems associated with the wide variation of electromyograph amplitude and background noise depending on the specific patient's muscles selected. A pulse-width modulated signal drives the DC motor, allowing closing and opening of the prosthesis. The relationship between the motor-driver signal and the actual hand-grasp force developed by the prosthesis was measured using a hand-held grip dynamometer. The resulting force was proportional to current for moderate values of current and then saturated. The motor torque, and, in turn, the force elicited, can be measured by sensing the current absorbed by the motor. Therefore, the grasp force can be opportunely limited or controlled. The cost of the only electronic and mechanical components of the electromyographically controlled hand was about US$50; other costs, such as the cost of labor to assemble the prosthesis and the production of adapters for patients, were not estimated.
