KMITL

Permanent URI for this communityhttps://dspace.kmitl.ac.th/handle/123456789/1

Browse

Search Results

Now showing 1 - 10 of 15

Some of the metrics are blocked by your
consent settings
ALIVE: An Agentic Longitudinal IDE-IPA Value Evaluation Framework for Innovation-Driven Enterprise Assessment
(2026-01-01)
Chaoraingern, Jutarut
;
Numsomran, Arjin
Agentic artificial intelligence architectures have demonstrated transformative potential for automating complex, multi-dimensional evaluation tasks; however, existing deployments remain predominantly stateless, incapable of accumulating institutional knowledge across evaluation cycles and therefore structurally unfit for longitudinal domains such as Impact Pathway Assessment and Social Return on Investment scoring, which inherently demand temporal coherence and precedent-informed judgment. This article presents ALIVE: an Agentic Longitudinal IDE-IPA Value Evaluation Framework, a sixth-generation (L6) multi-agent architecture addressing this limitation through the PACT Loop: Perceive normalizes inputs and retrieves episodic priors from ChromaDB; Analyze scores 18 dimensions in parallel via a shared rubric loaded once across all agents; Converge applies an auto-calibrating halting rule terminating on score stability, target achievement, or budget exhaustion; and Transfer extracts lessons via idempotent atom derivation, matches peer patterns, and broadcasts learnings, with longitudinal state in PostgreSQL and ChromaDB ensuring each cycle improves the next. Four agents (Funder, Company, Researcher, and Society), each instantiated as a role-specific skill context, are orchestrated by Claude via 21 stateless MCP tools on Railway.app without server-side API key. A self-evaluation loop scores each response against a structured rubric, triggering retry until a quality-gate is satisfied. ALIVE operationalizes the IDE-IPA Analyzer-Pro V2.0, a 100-point, 18-dimension rubric spanning Standard IDE Assessment (Part A), Research-Specific Assessment (Part B), Impact Pathway Logic (Part C), and SROI Assessment (Part D). Validation on thirty synthetic proposals across six industry domains against a three-expert panel demonstrates ICC<inf>21</inf>=0.922 (≥ 0.80 good-agreement threshold), funding decision accuracy of 86.7%, and sub-five-minute processing per proposal, establishing ALIVE as a scalable, self-improving infrastructure for longitudinal impact pathway assessment in research funding administration.
Some of the metrics are blocked by your
consent settings
Embedded Sensor Data Fusion and TinyML for Real-Time Remaining Useful Life Estimation of UAV Li Polymer Batteries
(2025-06-01)
Chaoraingern, Jutarut
;
Numsomran, Arjin
The accurate real-time estimation of the remaining useful life (RUL) of lithium-polymer (LiPo) batteries is a critical enabler for ensuring the safety, reliability, and operational efficiency of unmanned aerial vehicles (UAVs). Nevertheless, achieving such prognostics on resource-constrained embedded platforms remains a considerable technical challenge. This study proposes an end-to-end TinyML-based framework that integrates embedded sensor data fusion with an optimized feedforward neural network (FFNN) model for efficient RUL estimation under strict hardware limitations. The system collects voltage, discharge time, and capacity measurements through a lightweight data fusion pipeline and leverages the Edge Impulse platform with the EON™Compiler for model optimization. The trained model is deployed on a dual-core ARM Cortex-M0+ Raspberry Pi RP2040 microcontroller, communicating wirelessly with a LabVIEW-based visualization system for real-time monitoring. Experimental validation on an 80-gram UAV equipped with a 1100 mAh LiPo battery demonstrates a mean absolute error (MAE) of 3.46 cycles and a root mean squared error (RMSE) of 3.75 cycles. Model testing results show an overall accuracy of (Formula presented.), with a mean squared error (MSE) of 55.68, a mean absolute error (MAE) of 5.38, and a variance score of 0.99, indicating strong regression precision and robustness. Furthermore, the quantized (int8) version of the model achieves an inference latency of 2 ms, with memory utilization of only 1.2 KB RAM and 11 KB flash, confirming its suitability for real-time deployment on resource-constrained embedded devices. Overall, the proposed framework effectively demonstrates the feasibility of combining embedded sensor data fusion and TinyML to enable accurate, low-latency, and resource-efficient real-time RUL estimation for UAV battery health management.
Some of the metrics are blocked by your
consent settings
TinyML Speech Classification Embedded Ai Module for Hand Rehabilitation Device
(2025-01-01)
Numsomran, Arkorn
;
Numsomran, Arjin
;
Chaoraingern, Jutarut
Recent advancements in artificial intelligence and machine learning have significantly improved healthcare, especially in the development of assistive technologies for rehabilitation. This paper introduces a method for hand rehabilitation that utilizes the capabilities of tiny machine learning to enhance speech classification in a rehabilitation device. The proposed method employs a CNN model capable of classifying speech signals that are indicative of different hand movement patterns. Patients emit these speech signals during prescribed hand exercises, which are crucial for their rehabilitation process. The main focus of this study is on training and deploying a speech classification system that can work in the resource-limited environment of TinyML platforms. We detail the process of capturing speech data, preprocessing it, and extracting the most features relevant to different hand movements. Our results show that using TinyML to help with hand rehabilitation works. The method we came up with shows how TinyML could change the way rehabilitative devices are controlled, and it also shows us what personalized and easy-to-use rehabilitation tools might look like in the future.
Some of the metrics are blocked by your
consent settings
The Development of a Lithium-Ion Battery Analyzer for Data-Driven Analysis of Battery Performance
(2024-01-01)
Jaitrong, Kunanon
;
Seetisarn, Puvit
;
Numsomran, Arjin
;
Tipsuwanporn, Vittaya
;
Chaoraingern, Jutarut
As demand for higher energy density, faster charging, and longer lifespan increases, so too does the need for sophisticated analysis of battery performance. This study focuses on the development of a lithium-ion battery analyzer for data-driven battery performance analysis. The system comprises a hardware platform and a cloud-connected software infrastructure. The hardware utilizes an STM32 microcontroller integrated with an LTC6803 for precise battery parameter measurements, including voltage, current, and temperature. The charging circuit is an LTC-4162, allowing for controlled charging profiles. A Raspberry Pi, running Raspbian OS, acts as the central processing unit, collecting data from the microcontroller and displaying it on a user-friendly human machine interface developed using Node-RED. Crucially, the system leverages cloud connectivity for data storage, analysis, and remote monitoring. The system transmits measured data to a cloud platform, which offers scalable storage and facilitates the development of sophisticated data analysis algorithms. This cloud integration allows for real-time battery performance monitoring and visualization through a dedicated application. Additionally, the system incorporates a local data backup mechanism, ensuring data accessibility even during internet disruptions. The developed system provides a comprehensive platform for data-driven insights into battery health, enabling informed decisions regarding battery operation and maintenance.
Some of the metrics are blocked by your
consent settings
Real-Time Indoor Air Quality Index Prediction Using a Vacuum Cleaner Robot's AIoT Electronic Nose
(2023-01-01)
Chaoraingern, Jutarut
;
Tipsuwanporn, Vittaya
;
Numsomran, Arjin
To encourage the good health and well-being sustainable development goal, this article presents the design and implementation of real-time indoor air quality index (AQI) prediction using an artificial internet of things (AIoT) electronic nose integrated into a vacuum cleaner robot. The objective of the proposed method is to implement an effective embedded AIoT solution utilizing sensor fusion and the TinyML framework for the purpose of strengthening the environmental health system with suitable current technology. The high-accuracy sensor outputs of total volatile organic compounds (TVOC), humidity, equivalent carbon dioxide (eCO2), and PM2.5 gathered as the dataset are normalized in the data pre-processing state and utilized to create trained models using dense neural networks (DNN) deep learning algorithms. Tiny machine learning is responsible for neural network training, as it is capable of executing AI algorithms on embedded devices with extremely low power consumption and limited RAM and ROM resources. The testing results demonstrate that the predictive model performed well, with 99% accuracy for a maximum absolute regression error less than 15 and an 18.33 mean square error. The embedded device implementation uses Wio terminals with ARM Cortex-M4F microcontrollers for real-time indoor air quality index prediction and visualization. Experimental results demonstrating the average precision of the indoor AQI prediction were obtained at an average accuracy of over 98% with a computation time of 10 milliseconds and an acceptable usage of ROM and RAM resources of 8.5 KB and 1.1 KB, respectively, along with successive performances that satisfied web application data virtualization using the internet of things (IoT).
Some of the metrics are blocked by your
consent settings
The PIDmn Controller Tuning for Temperature and % RH Control in a Chamber Using Simulink Desktop Real-Time
(2023-01-01)
Boonseng, Nuengsatree
;
Numsomran, Arjin
;
Tipsuwanporn, Vittaya
;
Chaoraingern, Jutarut
This paper presents the development and implementation of a PID-{n}{m} controller tuning technique for temperature and relative humidity (% RH) control in a chamber using Simulink Desktop Real-Time. The proposed controller tuning method aims to achieve precise and robust control of temperature and % RH in environments where accurate environmental conditions are critical, such as laboratories, manufacturing facilities, and storage units. The PID-{n}{m} controller incorporates additional derivative, integral, and filter terms to enhance control performance. The design process involves analyzing the system dynamics, modeling the chamber's temperature and % RH response, and selecting appropriate control parameters. The tuning strategy takes into account the specific requirements of temperature and % RH control, such as fast response, a lower percent of overshoot, and steady-state accuracy. Simulink Desktop Real-Time is utilized for the real-time implementation of the control algorithm, providing a reliable and efficient platform for testing and validation. Experimental results demonstrate the effectiveness of the proposed methodology in achieving precise and stable temperature and % RH control.
Some of the metrics are blocked by your
consent settings
Event-Based LQR Control for Rotary Inverted Pendulum Using Wireless Networked Control System
(2023-01-01)
Thongsakul, Baitong
;
Numsomran, Arjin
;
Tipsuwanporn, Vittaya
;
Chaoraingern, Jutarut
Wireless communication introduces delays, packet losses, and data uncertainties, which can degrade the control's performance. To address these challenges, this paper presents an event-based Linear Quadratic Regulator (LQR) control approach for a rotary inverted pendulum system utilizing a Wireless Networked Control System (WNCS). To facilitate the control process, a WNCS is employed to transmit data between the rotary inverted pendulum and the controller via WiFi, a wireless network executing in the CYW43439 wireless chip of the Raspberry Pi Pico W development board. The wireless node consists of a sensor node, actuator node, access point node, and controller node, which transmit and receive data via the TCP/IP communication protocol. The event-triggering mechanism is integrated into the control loop. It allows the control algorithm to adaptively adjust the control update rate based on the system's dynamics and network conditions, ensuring efficient utilization of network resources while maintaining satisfactory control performance. The experimental results of hardware in the loop (HIL) in the MATLAB/Simulink real-time kernel show that the proposed event-based LQR control strategy can achieve balancing and improve resilience to network-induced disturbances, making it suitable for real-world applications where wireless network constraints are prevalent.
Some of the metrics are blocked by your
consent settings
Artificial Intelligence for the Classification of Plastic Waste Utilizing TinyML on Low-Cost Embedded Systems
(2023-01-01)
Chaoraingern, Jutarut
;
Tipsuwanporn, Vittaya
;
Numsomran, Arjin
BCG’s implementation of the economy makes Thailand more environmentally conscious. The consolidation policy encourages consumers to eliminate single-use plastics using the 3Rs. This article introduces a solution to reduce plastic waste drastically using artificial intelligence. Utilizing a low-cost Arducam Pico4ML embedded device and TinyML, a plastic waste classifying system prototype is developed for plastic bottle segregation. The grayscale image datasets of PET, HDPE plastic bottles, and unknown objects are adjusted in the image pre-processing state and utilized to create trained models using MobileNetV2 convolutional-based neural network algorithms. Effective feature extraction and model training are performed on the Edge Impulse platform, and the trained model is exported to an embedded device using the optimized compiler. A further RS485 Modbus communication protocol feature enables integration with a programmable logic controller (PLC). The validation results of the trained model indicate a classification performance of 100% accuracy. Based on the average precision results, it is notable that the trained model can recognize the most common waste with an average accuracy of over 90%. The minimum classification rate of the MobileNetV2 quantized model is 249 milliseconds. It is also implemented in low-cost embedded devices for real-time plastic waste classification using fewer processing resources (185.4K ROM and 88K RAM). The findings exhibit sequential contributions that satisfy the criteria for classifying plastic bottles and the machine’s integration capacity. These outcomes are anticipated to foster social shifts in behavior and enhance public awareness about plastic waste management.
Some of the metrics are blocked by your
consent settings
Modified adaptive sliding mode control for trajectory tracking of mini-drone quadcopter unmanned aerial vehicle
(2020-10-01)
Chaoraingern, Jutarut
;
Tipsuwanporn, Vittaya
;
Numsomran, Arjin
Unmanned Aerial Vehicles are among the most widely attracting interest, especially in the applications of transportation, inspection, and surveillance. The great mechanisms for motion control are set to become a vital factor in performing the robust and accurate stabilized flight relied on perturbations and disturbances. This paper presents modified adaptive sliding mode control for trajectory tracking of mini-drone quadcopter unmanned aerial vehicles, which aims at demonstrating the effectiveness of nonlinear adaptive control strategy for achieving the desired performance of the mini-drone quadcopter system. Besides providing mathematical modeling and nonlinear dynamic characteristic details of mini-drone quadcopter actuated system, the modified adaptive sliding mode algorithm is developed using adaptation law based on Lyapunov stability approach then applied on the attitude loop and the altitude loop control system so that the nonlinear adaptive behavior of the controller enables the compensation of disturbances and parameter perturbations. The effectiveness validations of the proposed control technique compared with the traditional approach are performed through the Matlab simulation. The results have been illustrated that the modified adaptive sliding mode control can decrease the error performance indexes to the minimum ISE at 1.041 m2 and the zero percentage of overshoot while enables excellent stability and robustness even in the presence of parameter perturbations and disturbance.
Some of the metrics are blocked by your
consent settings
Efficient fractional order reference model of adaptive controller design for multi-input multi-output thermal system
(2020-01-01)
Chaoraingern, Jutarut
;
Tipsuwanporn, Vittaya
;
Numsomran, Arjin
The diverse control techniques have been combined with fractional calculus to enhance the control system performance. This paper presents an efficient fractional-order model reference adaptive controller (FOMRAC) design, which aims to demonstrate the solution for temperature reference tracking and cross-coupling rejection in the multi-input multi-output thermal system as well as cognizing of power consumption saving constraints. The mathematical modeling, nonlinear dynamic characteristic details, and system identification of the thermal system are described while the fractional-order controller combined with a model reference adaptive control (FOMRAC) based on MIT rule is developed so that to create the nonlinear adaptive mechanism which enables the excellent performance to control the multi-input multi-output thermal system. Likewise, a decoupling compensator is constructed to remunerate the effect of the cross-coupling interaction. The validation of the proposed control scheme is performed through the Matlab simulation and the experiment on the multi-input multi-output thermal system. The results illustrated the FOMRAC technique in which the controller's adjustable parameters can provide efficiency stability and performance to minimize the settling time and percent overshoot of the control system response. Besides, the analysis of the power consumption in the control system is addressed to reinforce the useful ability of the proposed method compared with the integral-order model reference adaptive controller (IOMRAC) and the traditional PID controller. The results revealed that the proposed FOMRAC technique exhibited much better than other methods because of the effective optimization of adaptive gain mechanism and fractional-order operators.

KMITL

Browse

Filters

Settings

Sort By

Results per page

Search Results