A. Sumarudin

Reinforcement Learning (RL) applications are rapidly expanding across various fields, including robotics, autonomous vehicles, drones, communication systems, and control. However, RL computations are inherently complex and resource-intensive due to dynamic state spaces and real-time processing demands. This paper proposes a novel RL accelerator hardware architecture for edge computing, capable of supporting both inference and training modes. The design, tailored for Deep Q-Network (DQN) with Q-values computed using an Artificial Neural Network (ANN), employs hardware sharing and approximate activation functions and multipliers to enhance efficiency. Shared processing modules minimize chip size and power consumption, achieving an area efficiency of 44.25 KIPS/1KLUTs and 14.65 KIPS/1KLUTs. Implementation results demonstrate state processing times of 1.43 μs for inference and 4.31 μs for training. Deployed on an Ultra96 FPGA at 70 MHz, the system delivers 671× and 287× speedups for inference and training, respectively, with a DQN accelerator power consumption of just 13 mW. This architecture is highly suitable for autonomous systems in edge computing environments, addressing resource limitations and enhancing computational efficiency.

The integration of AI into various sectors, including agriculture, has been advancing significantly. Implementing AI in the context of IoT and edge AI presents challenges due to resource limitations. Current climate changes affect planting strategies, pest management, and harvest timing. This study explores an SVM-based machine-learning model with multiple kernels to classify weather conditions as rainy or clear. The research includes two phases: model training on a PC-based system and model deployment on an edge AI device. The training phase includes preprocessing with PCA and fine-tuning of parameters, such as kernel types (linear, polynomial, sigmoid, and RBF), C and gamma. The development phase involves deploying the model on an ESP32, where execution time and power consumption are evaluated. The results show that the SVM model with an RBF kernel, C of 0.1, and gamma of 1 achieves a precision of 79.37%. Inference on the ESP32 yields an average execution time of 35.5 ms and a power consumption of 66 mA, showing a 202-fold reduction in power usage compared to the PC-based system and a 59-fold increase in execution time. This reduced power consumption supports the feasibility of edge AI for climate-based agricultural applications, enabling effective rainfall prediction. The findings contribute to the development of precision agriculture by providing insights into climate prediction, which can inform planting decisions, pest management, and harvest timing, thereby advancing the application of edge AI in response to global climate change.

One of the city problems is public transportation. The main public transportation in a city of indonesia is "angkot". The many types of angkot in Indonesia, makes it difficult to monitor the existence of these angkot. This problem is included to key element of smart city. In this study, we built a public transportation tracking system to support smart city. The public transportation tracking system that our propose consist of perform which is integrated angkot tracer between device, web server and mobile application. Devices are installed on public transportation, then web servers are used to monitor related services and cellular applications are used for passengers to get information on angkot locations. From the results of this study, it is found that the device able to read coordinate position with a error rate of 1.4 meters and send to the server with "get" method with json parsing to be monitored. Moreover, the mobile apps can display coordinates of the angkot by shown angkot id, distance, duration and route from angkot. This application can provide solutions for angkot monitoring to reduce public transportation violations and passengers can know position of angkot in realtime.

Indonesian agriculture has great potensial for development. Agriculture a lot yet based on data collection for soil or plant, data soil can use for analys soil fertility. We propose e-agriculture system for monitoring soil. This system can monitoring soil status. Monitoring system based on wireless sensor mote that sensing soil status. Sensor monitoring utilize soil moisture, humidity and temperature. System monitoring design with mote based on microcontroler and xbee connection. Data sensing send to gateway with star topology with one gateway. Gateway utilize with mini personal computer and connect to xbee cordinator mode. On gateway, gateway include apache server for store data based on My-SQL. System web base with YII framework. System done implementation and can show soil status real time. Result the system can connection other mote 40 meters and mote lifetime 7 hours and minimum voltage 7 volt. The system can help famer for monitoring soil and farmer can making decision for treatment soil based on data. It can improve the quality in agricultural production and would decrease the management and farming costs.

We proposed an easily configured Wi-Fi wireless sensor network for mesh topology using OLSR (optimized link state protocol). The system is capable to connect between motes with mesh topology. In order to support high data bitrate, multi applications, multi platform, and secure data connection, we implement the system using Raspberry Pi(R) with ARMv7(R) Quad Cores and Wi-Fi interface. For easy of configuration and operation, the system utilized Linux based Raspbian kernel 3.18 operating system. The system has been successfully implemented and verified for real-time soil monitoring system. The system implementation only utilized 1% of processor resources and very responsive for topology regenerating. The QoS performance shows the average jitter is 3.87 ms, the throughput is 8.5 Mbps and the Packet loss is 0.019%. In term of range, the system can reach 175 Meter. The result shows that the proposed WSN can be used for high bit-rate Wi-Fi based on application.

This paper proposed a reconfigurable and flexible SoC architecture for various applications of wireless sensor network (WSN). The hardware-software system reconfiguration can be easy done because the systems utilize FPGA based SoC as a platform. The system also supports various network topologies and interfaces. The supported interface consists of I2C, SPI, USART, etc. The main processing system of SoC is based on ARM v7 processor that has high computing capability. The system utilizes Ubuntu 12.04 LTS as operating system. As consequences, we can easily interface the system with many external devices and create various program applications, such as monitoring system, sensor data warehouse, etc. The system has been successfully implemented with communication interface and PH, temperature, and humidity sensors. The system also has been verified using designed monitoring application for PH, temperature and humidity data that are transferred into gateway.