| 1. |
- Calatrava Nicolás, Francisco M., 1997-, et al.
(author)
-
Light Residual Network for Human Activity Recognition using Wearable Sensor Data
- 2023
-
In: IEEE Sensors Letters. - : IEEE. - 2475-1472. ; 7:10
-
Journal article (peer-reviewed)abstract
- This letter addresses the problem of human activity recognition (HAR) of people wearing inertial sensors using data from the UCI-HAR dataset. We propose a light residual network, which obtains an F1-Score of 97.6% that outperforms previous works, while drastically reducing the number of parameters by a factor of 15, and thus the training complexity. In addition, we propose a new benchmark based on leave-one (person)-out cross-validation to standardize and unify future classifications on the same dataset, and to increase reliability and fairness in the comparisons.
|
|
| 2. |
- Fhager, Lars Ohlsson, et al.
(author)
-
Pulsed Millimeter Wave Radar for Hand Gesture Sensing and Classification
- 2019
-
In: IEEE Sensors Letters. - 2475-1472. ; 3:12
-
Journal article (peer-reviewed)abstract
- A pulsed millimeter wave radar operating at a frame rate of 144 Hz is utilized to record 2160 scattering signatures of 12 generic hand gestures. Gesture recognition is achieved by machine learning, utilizing transfer learning on a pretrained convolutional neural network. This yields excellent classification results with a validation accuracy of 99.5%, based on a 60% training versus 40% validation split. The corresponding confusion matrix is also presented, showing a high level of classification orthogonality between the tested gestures. This is the first demonstration where data from a pulsed millimeter wave radar is used for gesture recognition by machine learning. It proves that the range-time envelope representation of high frame-rate data from a pulsed radar is suitable for hand gesture recognition. Further improvements are expected for more complex detection schemes and tailored neural networks.
|
|
| 3. |
- Ivanov, Ivan, et al.
(author)
-
Selective Low-Temperature Hydrogen Catalytic Sensor
- 2022
-
In: IEEE SENSORS LETTERS. - : Institute of Electrical and Electronics Engineers (IEEE). - 2475-1472. ; 6:5
-
Journal article (peer-reviewed)abstract
- The research demonstrates for the first time the high selectivity of catalytic hydrogen sensors to other hydrocarbons (methane, propane, hexane, butane, ethane, and ethylene) at a temperature less than 70 degrees C. Two circuits were used to measure the response of the sensors: a Wheatstone bridge circuit and a divider circuit. Hydrogen measurement was conducted within a temperature range of 66-130 degrees C. The sensors exhibited high sensitivity (25-35 mV/%) and low power consumption (approximately 8.6 mW). The Wheatstone bridge circuit was observed to have the maximum value of selectivity and sensitivity.
|
|
| 4. |
- Johansson, Ted, 1959-, et al.
(author)
-
Improving angle-of-view for a 1-D sensing application by using a 2-D optical sensor in "cylindrical" mode
- 2021
-
In: IEEE Sensors Letters. - : IEEE. - 2475-1472. ; 5:10
-
Journal article (peer-reviewed)abstract
- To further develop a low-power low-cost optical motion detector for use with traffic detection under dark and daylight conditions, we have developed and verified a procedure to use a Near Sensor Image Processing (NSIP) programmable 2-D optical sensor in a "1-D mode" to achieve the effect of using a cylindrical lens, thus improving the angle-of-view (AOV), the sensitivity, and usefulness of the sensor. Using an existing 256 x 256 element sensor in an innovative way, the AOV was increased from 0.4 to 21.3 in the vertical direction while also improving the sensitivity. The details of the sensor hardware architecture are described in detail and pseudo code for programming the sensor is discussed. The results were used to demonstrate the extraction of Local Extreme Points (LEPs) used for Time-To-Impact (TTI) calculations to estimate the speed of an approaching vehicle.
|
|
| 5. |
- Johansson, Ted, et al.
(author)
-
Improving Angle-of-View for a 1-D Sensing Application by Using a 2-D Optical Sensor in "Cylindrical" Mode
- 2021
-
In: IEEE SENSORS LETTERS. - : Institute of Electrical and Electronics Engineers (IEEE). - 2475-1472. ; 5:10
-
Journal article (peer-reviewed)abstract
- To further develop a low-power, low-cost optical motion detector for use with traffic detection under dark and daylight conditions, we have developed and verified a procedure to use a near-sensor image processing programmable 2-D optical sensor in a "1-D mode" to achieve the effect of using a cylindrical lens, thus improving the angle-of-view (AOV), the sensitivity, and usefulness of the sensor. Using an existing 256 x 256 element sensor in an innovative way, the AOV was increased from 0.4. to 21.3. in the vertical direction while also improving the sensitivity. The details of the sensor hardware architecture are described in detail and pseudo-code for programming the sensor is discussed. The results were used to demonstrate the extraction of local extreme points used for time-to-impact calculations to estimate the speed of an approaching vehicle.
|
|
| 6. |
- Johansson, Ted, et al.
(author)
-
Low-Power Optical Sensor for Traffic Detection
- 2020
-
In: IEEE Sensors Letters. - : Institute of Electrical and Electronics Engineers Inc.. - 2475-1472. ; 4:5
-
Journal article (peer-reviewed)abstract
- A CMOS sensor chip was used, together with an Arduino microcontroller, to create and verify a low-power low-cost optical motion detector for use in traffic detection under dark and daylight conditions. The chip can sense object features with very high dynamic range. On-chip near sensor image processing was used to reduce the data to be transferred to a host computer. A method using local extrema point detection was used to estimate motion through time-to-impact (TTI). Sensor data from the headlights of an approaching/passing car were used to extract TTI values similar to estimations from distance and speed of the object. The method can be used for detection of approaching objects to switch on streetlights (dark conditions) or sensors for traffic lights instead of magnetic sensors in the streets or conventional cameras (dark and daylight conditions). A sensor with a microcontroller operating at low clock frequency will consume less than 30 mW in this application.
|
|
| 7. |
- Karlsson, Rickard, 1970-, et al.
(author)
-
Speed Estimation From Vibrations Using a Deep Learning CNN Approach
- 2021
-
In: IEEE Sensors Letter. - : Institute of Electrical and Electronics Engineers (IEEE). - 2475-1472 .- 2475-1472. ; 5:3
-
Journal article (other academic/artistic)abstract
- A novel method for accurate speed estimation of a vehicle using a deep learning convolutional neural network (CNN), with accelerometer and gyroscope measurements as input, is presented. It does not suffer from the fundamental drift problem present in all dead reckoning methods, and yet yields about 2 m/s in accuracy. Efficient drift-free vehicle speed estimates are essential in many automotive applications, where internal wheel speed sensors or GPS are unavailable. Using extensive experimental data, the proposed CNN method is compared to an existing frequency analysis method. The proposed method is shown to perform significantly better, particularly during low speed and rapid speed changes where the frequency method struggles.
|
|
| 8. |
|
|
| 9. |
- McGurk, Colin, et al.
(author)
-
A Practical Guide for Noise Characterisation of Back Pressure Sensors: Towards Digital Twin for an Industrial High Horse Power Engine Test Cell
- 2024
-
In: IEEE Sensors Letters. - : IEEE. - 2475-1472. ; 8:6
-
Journal article (peer-reviewed)abstract
- Extensive testing is crucial to ensure exhaust emission compliance for high-horsepower (H-hp) commercial engines. Back pressure sensors, integral to the exhaust system, are prone to producing noisy measurements due to the turbulent nature of the exhaust gas flow and other testing inaccuracies, mandating the use of a noise reduction filter. The time-consuming process of filter tuning, required to remove excessive process/measurement noise, often involves trialand-error. This practice, entailing numerous experiments with live engines, results in high financial costs and emissions due to challenges in extracting the ground truth signal from noisy measurements. Developing a digital twin (DT) of this system is proposed to expedite filter tuning, hence reducing cost and emissions output. Creating such a DT necessitates a method of back pressure sensor noise classification to accurately simulate the signal. This letter introduces a step-by-step procedure for characterising the H-hp engine back pressure sensor through statistical measures, leading to the development of the DT. This letter demonstrates the potential of this approach in an industrial case study, showcasing its viability for application in engine test-bed facilities and across industries. The economic calculation estimates a potential £184,400 reduction in diesel fuel costs and 321,600 kg of CO2 emissions by tuning the filter through the DT compared to current industrial practice
|
|
| 10. |
- Skog, Isaac, 1981-
(author)
-
Nonintrusive Elevator System Fault Detection Using Learned Traffic Patterns
- 2020
-
In: IEEE Sensors Letters. - : Institute of Electrical and Electronics Engineers (IEEE). - 2475-1472. ; 4:11
-
Journal article (peer-reviewed)abstract
- A new method for nonintrusive elevator fault detection is presented. A computationally efficient algorithm for implementing the method is also proposed. The method is employed to detect when the elevator has been stationary for an unusually long period of time compared to historical traffic load patterns. This information can be used for fault detection but also indirectly to monitor the condition of the doors. The traffic load on the elevator is modeled as a nonhomogeneous Poisson process, and a generalized linear model is used to describe how the intensity of the process varies over time. A statistical hypothesis test is then used to determine if the elevator has been stationary for an unusually long time. The application of the proposed method is illustrated by an example where the detected faults are compared with the elevator service log. All faults were detected long before the service company was notified by the facility owner. Furthermore, based on the evaluation of 30 weeks of data, the method achieves a precision of 0.82 at a recall probability of 0.80.
|
|
