Activity Recognition using Multi-Class Classification inside an Educational Building

Abstrak –This study aims to: (1) Design a prototype of Bird Pest Repellent on Rice Plants Using PIR and ESP8266 Sensors in Matekko Bulukumba Village. (2) Knowing the implementation of the prototype of Bird Pest Repellent on Rice Plants Using PIR and ESP8266 Sensors in Matekko Bulukumba Village. This study uses a prototype research method where the stages are as include collection of needs, make prototype, prototype evaluation, encoding system, System testing, System evaluation & Using the system. Based on the results of this study, it can be concluded that: (1) The prototype of Bird Pest Repellent on Rice Plants was designed using 3 components, namely 3 PIR (Passive Infra Red) Sensors, 1 Esp8266 and 3 Servo Motors. This prototype of Bird Pest Repellent on Rice Plants uses a plywood board as a place to place the components of the tool which has a width of 47 centimeters and a length of 50 centimeters. The PIR (Passive Infra Red) sensor will be placed at the end of each square-shaped plywood board. ESP8266 will be placed in the middle of the Servo Motor and the Servo Motor will be placed at the end of the upper plywood board. (2) The implementation of the bird repellent prototype on rice plants is able to detect the presence (bird pests) by means of a servo motor moves the rope and provides notifications to the blynk application.

Passive infrared (PIR) sensors are the most popular deployed sensors in building lighting control for individual presence detection. However, PIR sensors are motion detectors in nature, responding only to incident radiation variation, which lead to false negative detections, inaccurate occupancy estimation, and uncomfortable lighting swings, short lifetime of the equipment, and waste of energy. In this study, a shutter driven by a Lavet motor PIR (LAMPIR) sensor is developed for presence detection for both stationary and moving occupants. Building off our previous work on chopped PIR (C-PIR) and rotationally-chopped PIR (Ro-PIR) sensors, Lavet motor, a single-phase electro-mechanical vibrator, is introduced, which has many advantages over traditional servo motors and stepper motors in terms of power consumption, size, weight and noise level. Driven by pulsed signal from a microcontroller unit (MCU), the electro-mechanical vibrator drives a semi-transparent long-wave infrared (LWIR) optical shutter to shutter the field of view (FOV) of a PIR sensor periodically. Output voltage generated by a LAMPIR senor for occupied and unoccupied scenarios can be monitored and analyzed to identify presence accurately. Parametric studies are conducted to find the optimal setting of driving signal frequency, shutter width and shuttering period. The LAMPIR sensor reaches an accuracy of 100% for detecting stationary occupants up to a range of 4.5 m and moving occupants up to a range of 10 m, which improves the detection range of both C-PIR and Ro-PIR sensors (4.0 m for stationary and 8.0 m for moving occupancy detection). LAMPIR has a FOV of 90° in horizontal and 100° in vertical, which is reasonable for most applications. For a 17-hour-long presence detection test, LAMPIR can reach an accuracy of 93.52% to classify unoccupied, stationary and moving occupant scenarios. More importantly, the average power consumption of LAMPIR is 0.19 W, which is 82% less than that of the C-PIR sensor and 89% less than that of the Ro-PIR sensor.

Deep learning has shown to be capable of learning features from complex datasets such as in the case of biometric authentication. Biometric authentication relies on unique biological qualities to verify a user's identity against a group of potential adversaries for security purposes. In this paper, a deep learning dependent biometric authentication system based on chest motion data captured by a passive infrared (PIR) sensor, CM-PIR, is proposed. PIR sensors are cheap, commercial-off-the-shelf (COTS) components that are dependent on motion across its field of view (FoV) for accurate detection of human subjects. Thus, CM-PIR utilizes deep learning for the accurate detection of stationary human subjects, as well as for biometric authentication. CM-PIR collects chest motion data from sixteen human subjects in nine different home locations. Coefficients of Fourier transform (FFT), discrete wavelet transform (DWT), and the absolute value of an acceleration filter calculated from the raw voltage PIR values were selected as the optimal features for input to the deep learning models for biometric authentication. Preprocessing of these features included a threshold voltage range, normalization, and finally principal component analysis (PCA). CM-PIR is 94% accurate in stationary human presence detection and 75% accurate for biometric authentication using a RNN with a 90 second window size. This work provides a promising solution for stationary human detection and biometric authentication using a PIR sensor in a real-world setting.

Research has been carried out aimed at:(1) to produce a home security system using a PIR (passive infrared) sensor with voice warnings and telegram notifications based on the MCU node microcontroller. (2) to determine the performance of the resulting home security system. The design uses a node mcu microcontroller as the main controller. The design uses a motion sensor (PIR) to detect motion that occurs inside the house. The design uses an internet modem to send telegram notifications to homeowners. The speed of communication between the microcontroller and Telegram depends on the internet network connection. Sending messages via telegram to the mcu node, it is also assumed that electricity and internet are in a stable condition. A home security system using a PIR (passive infrared) sensor with sound warnings and telegram notifications based on the MCU node microcontroller can be generated. Testing of each component (power supply, PIR sensor, MCU node microcontroller, MP3 player module, limit switch sensor and ESP8266) indicates that each part is functioning and working properly. The performance of the resulting home security system is very good. After the system is activated, any movement detected in the living room and/or terrace of the house as well as windows and/or doors that are detected are open, the system will respond by emitting a sound through the speaker in the form of a warning or informing that a house break-in has occurred while sending a chat telegram to the user.

Human activity recognition is one of the most challenging research areas in human-machine interaction. Activity recognition by wearable sensors produces significant information about a person's functional ability and lifestyle. In this paper, we present an inertial sensors-based approach for activity and postural transition recognition. Efficient features are extracted from the sensor data. Then, the features are normalized using Z-score followed by Generalized Discriminant Analysis (GDA) to make them more robust. Finally, the features are trained with a Recurrent Neural Network (RNN) for activity and postural transition recognition. The proposed method recognizes 12 activities with an overall accuracy of 97%. In the activity recognition system, the multimodal input data is collected from only one accelerometer and one gyroscope sensor embedded in the user's smartphone. The proposed system can contribute in research fields such as human-robot interaction for practical applications such as eldercare in smart homes.

Adaptive Bayesian inference system for recognition of walking activities and prediction of gait events using wearable sensors

Food is necessary to support daily life and in pets like cats. Pet feeders are usually carried out routinely by pet owners by taking the time to stay at home and feed the pets. With activities requiring animal owners to be outside the home for long periods, it is necessary to design a prototype that can automatically assist the process of providing food, especially dry food, to pets. So pet feeders can be carried out even though the owner is not at home and does not have spare time. The system was built using the C language related to the Code Vision tools that support the hex file compiler into the microcontroller. The research method used is the experimental method, namely conducting trials (trial and error) directly on the research object and the comparative testing method for testing the built system. The auto pet feeder prototype uses the ATMega8535 IC by utilizing a PIR (Passive InfraRed) sensor to detect movement around the animal bowl. The PIR sensor detects well as expected, and it has an effective detection range of up to 5 meters. Not only detects humans, but the PIR sensor can also detect other living things. The PIR sensor will send a signal to the microcontroller to open the valve from the food measuring station, which is driven by a DC motor based on the detected motion to drop food into the animal's bowl. PhotoDioda controls the amount of food falling into the animal's bowl. The test results of this tool show that at a distance of 0–6 meters, the PIR sensor can provide feedback or detect animal movements well.

Human Activity Recognition (HAR) based on accelerometer has become an important mobile application. Activity recognition however depends on X, Y, Z the Cartesian coordinate parameters. There are several approaches for activity recognition. Popular methods of activity recognition using accelerometer reading includes machine learning approach, rule based data mining approach, fuzzy inference approach etc. This paper compares activity recognition based on temporal pattern mining and ANFIS method for wearable sensor accelerometer and mobile accelerometer readings. Though the existing activity recognition using body worn accelerometer gives better accuracy it is found to be costly and consume more power. Hence this paper proposes an ANFIS based activity recognition with the available accelerometer in mobile phone.

Activity recognition is one of the preliminary steps in designing and implementing assistive services in smart homes. Such services help identify abnormality or automate events generated while occupants do, as well as intend to do, their desired activities of daily living (ADLs) inside a smart home environment. However, most of the existing systems are applied for single-resident homes. Multiple people living together create additional complexity in modeling numbers of overlapping and concurrent activities. In this paper, we introduce a hybrid mechanism between ontology-based and unsupervised machine learning strategies in creating activity models used for activity recognition in the context of multi-resident homes. On comparing with related data-driven approaches, the proposed technique is technically and practically scalable to real-world scenarios due to fast training time and easy implementation. An average activity recognition rate of 95.83% on CASAS Spring dataset was achieved and the run time per recognition operation was measured as 12.86 ms.

The aim of this research is to develop Atwood machine propswith sensor-based PIR (Passive Infrared) to determine the acceleration value of earth gravity. This type of research is an experiment conducted in the physics laboratory of Universitas Indraprasta PGRI. Atwood machine was developed by using PIR (Passive Infrared) sensors to detect the movement of the objects used. Based on the trialof the result of the data analysis experiment tool, the acceleration of gravity was obtained (9.87 ± 0.08) m/s2. Whereas the value of acceleration of gravity in the literature is 9,80665 m/s2. The obtained results shows that the acceleration of gravity obtained in this research is not much different from the acceleration value of gravity found in the literature so that Atwood machine tool that is developed by using PIR sensor is recommended to be one alternative tool in determining the acceleration of earth gravity.

To enable passive infrared (PIR) sensors to detect stationary occupants, we recently invented a synchronized low energy electronically chopped PIR (SLEEPIR) sensor consisting of PIR sensor, a polymer dispersed liquid crystal (PDLC) infrared shutter and a driving circuit. By electrically chopping the long-wave infrared (8- <inline-formula xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink"> <tex-math notation="LaTeX">$12~\mu \text{m}$ </tex-math></inline-formula> ) radiation received by the PIR sensor, the SLEEPIR sensor can detection both stationary and moving occupants. In this paper, we developed a theoretic model to better understand the output of the SLEEPIR) sensor, and thus optimize its fabrication process and detection performance. The model indicated that its performance is mainly depending on the effective modulation and response time of the PDLC shutter in addition to the time constants of the PIR sensor. We empirically found that the effective modulation and response time are largely dependent on the frequency and amplitude of the driving signal, as well as the mass ratio, cell gap and cooling rate of the PDLC shutter. In the end we obtained optimized parameters that can produce the maximal SLEEPIR output, and experimentally validated our obtained optimal parameters and fabrication conditions.

Traffic lights at crossroads help regulate the flow of vehicles so they can run smoothly and avoid congestion. One of the causes of traffic congestion is a traffic control light system that is not optimal because it still uses a fixed timer system, so the lights' duration is less efficient in its application. When the green light is on, the streets are deserted. But there is a traffic jam on another road. Therefore the concept of traffic light timing needs to be developed to get efficient timing. This study aims to create a traffic light control system that uses a PIR (Passive Infrared) motion sensor and an ultrasonic sensor based on the Arduino Mega2560 microcontroller based on the congestion level. The method used in this traffic control prototype is an experimental method of reading vehicle objects mounted with PIR sensors and ultrasonic sensors. This system can detect the movement of vehicles approaching a traffic light, measure the distance between the car and the traffic light, and set the timing of the light signal according to the level of congestion detected. This system has been tested using vehicle movement simulations brought closer to the PIR and ultrasonic sensors at different distances. The test results show the system can work properly and produce the light signal according to the detected traffic conditions. For its application to traffic lights in the future, it is feasible to help the traffic unit.

Advancements in the implementation of quantum hardware have enabled the acquisition of data that are intractable for emulation with classical computers. The integration of classical machine learning (ML) algorithms with these data holds potential for unveiling obscure patterns. Although this hybrid approach extends the class of efficiently solvable problems compared to using only classical computers, this approach has been only realized for solving restricted problems because of the prevalence of noise in current quantum computers. Here, we extend the applicability of the hybrid approach to problems of interest in many-body physics, such as predicting the properties of the ground state of a given Hamiltonian and classifying quantum phases. By performing experiments with various error-reducing procedures on superconducting quantum hardware with 127 qubits, we managed to acquire refined data from the quantum computer. This enabled us to demonstrate the successful implementation of theoretically suggested classical ML algorithms for systems with up to 44 qubits. Our results verify the scalability and effectiveness of the classical ML algorithms for processing quantum experimental data.

Smart trash cans use a proximity sensor (ultrasonic) to detect the height of the trash, the PIR (Passive Infra Red) sensor to detect human presence and Arduino as the data processing center. This trash can also has an LED that functions as an indicator and the buzzer functions as an alarm when the trash can is full. If the garbage reaches half of the trash, the LED will turn yellow, if the garbage reaches a height of 4cm from the trash lid, then the LED will turn red and so the trash can will be locked even though the PIR sensor detects humans. Buzzer will live to attract the attention of cleaning officers to dispose of trash. To dispose of garbage, the clerk needs to shift the open and close trash button so that the lid will open. The author hopes that this smart trash can can increase public awareness to dispose of garbage in its place. The results of study show that PIR sensors work by well in detecting the presence of humans and ultrasonic sensors in detecting the height of garbage are not accurate, so that if the trash if the trash can is empty, sometimes the garbage is minus, whereasif the garbage is &lt; 3.5 cm, the sensor fails detect it.

Monitoring daily activities is essential for home service robots to take care of the older adults who live alone in their homes. In this article, we proposed a sound-based human activity monitoring (SoHAM) framework by recognizing sound events in a home environment. First, the method of context-aware sound event recognition (CoSER) is developed, which uses contextual information to disambiguate sound events. The locational context of sound events is estimated by fusing the data from the distributed passive infrared (PIR) sensors deployed in the home. A two-level dynamic Bayesian network (DBN) is used to model the intratemporal and intertemporal constraints between the context and the sound events. Second, dynamic sliding time window-based human action recognition (DTW-HaR) is developed to estimate active sound event segments with their labels and durations, then infer actions and their durations. Finally, a conditional random field (CRF) model is proposed to predict human activities based on the recognized action, location, and time. We conducted experiments in our robot-integrated smart home (RiSH) testbed to evaluate the proposed framework. The obtained results show the effectiveness and accuracy of CoSER, action recognition, and human activity monitoring. <italic xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">Note to Practitioners</i> —This article is motivated by the goal to develop companion robots that can assist older adults living alone. Among many capabilities, monitoring human daily activities is an essential one for such robots. Though computer vision or wearable sensors-based methods have been developed by other researchers, they are not practical due to the privacy concern and intrusiveness. Sound-based daily activity recognition can address these concerns and offer a viable solution. In this regard, our proposed method adopts microphones on the robot and a small set of motion sensors distributed in the home. The proposed theoretical framework was tested in a small-scale mock-up apartment with promising results. Before such companion robots can be deployed to real homes for elderly care, there is a need to improve the robustness of the algorithms. More thorough tests in various realistic home environments should be conducted to fully evaluate the performance of the robots. In addition, privacy concern related to audio capture should be further mitigated.