Adha Imam Cahyadi

UAV obstacle avoidance using potential field under dynamic environment

In this paper, the potential field principle is applied to several UAVs (Unmanned Aerial Vehicles) for optimal path planning. Each UAV has its own goals and it is used the attractive potential field to reach the goals. On the contrary, each UAV is considered as obstacle for other UAVs that must be avoided. In this research, there are two types of obstacles, i.e the static and dynamic. The repulsive potential field principle is used to avoid for both static and dynamic obstacles. The whole method is implemented into Parrot AR Drone 2.0 Quadcopter model of UAV and simulated in Gazebo Simulator by Robot Operating System (ROS). The results of this research are to control the thrust of quadcopter so that it is in flying position, the value of value of roll (α) or pitch (β) must set to not equal or approaching 90° or -90° and not between -180° and 180° or between -90° and 90°. Based on the dynamic obstacle performance test with parameter tuning, the optimal avoidance is when the η value is 7.8 while when in static and dynamic test with parameter tuning, the optimal avoidance is when the η value is 7.9 noted by the fastest time and the shortest path.

3D OBJECT MODELING USING DATA FUSION FROM LASER SENSOR ON QUADROTOR

Environmental modeling of a robot which is needed for robot navigation and path planning is in the form of planar or 2D modeling. Several previous researchers have used laser sensor to model 2D obstacles because it has data accuracy for navigation. It is in a planar shape and implemented in quadrotor thus the obstacle modeling formed is in 2D. The problem in a 2D environment is that the path planning and navigation of the robot requires a considerable time because the robot stops at loca minima and attempts to find other paths in the dimension. The problem can be solved by using combined data taken from the laser sensor. The combination of the data uses several algorithms such as graph theory and vector field histogram algorithms. Therefore, this paper presents the combination of the algorithms to model a 3D environment. By using this model, the quadrotor is able to avoid loca minima.

Path planning of decentralized multi-quadrotor based on fuzzy-cell decomposition algorithm

The paper aims to present a design algorithm for multi quadrotor lanes in order to move towards the goal quickly and avoid obstacles in an area with obstacles. There are several problems in path planning including how to get to the goal position quickly and avoid static and dynamic obstacles. To overcome the problem, therefore, the paper presents fuzzy logic algorithm and fuzzy cell decomposition algorithm. Fuzzy logic algorithm is one of the artificial intelligence algorithms which can be applied to robot path planning that is able to detect static and dynamic obstacles. Cell decomposition algorithm is an algorithm of graph theory used to make a robot path map. By using the two algorithms the robot is able to get to the goal position and avoid obstacles but it takes a considerable time because they are able to find the shortest path. Therefore, this paper describes a modification of the algorithms by adding a potential field algorithm used to provide weight values on the map applied for each quadrotor by...

Passive balancing of battery lithium polymer using shunt resistor circuit method

In this study, the system was designed for monitoring the voltage condition of each battery’s cell. The method used in this study is by using passive shunt resistor balancing method. An electronic circuit was designed in order to balance the value of the able voltage at the battery cells using resistors and then to remove the excess voltage. The result showed that the electrical circuit was capable to balance the voltage of each cell. Based on experiment with several values of load, it is can concluded that 0.1C rate of discharge has the best performance, because it does not affect battery voltage characteristic significantly which leads to better sensor reads.

The comparison of RBF NN and BPNN for SOC estimation of LiFePO4 battery

State of Charge (SOC) defined as the percentage of remaining capacity relative to the maximum capacity of the battery. In Battery Management Systems (BMS), SOC is an important variable. In this paper will describe comparison between Backpropagation Neural Networks (BPNN) and Radial Basis Function Neural Network (RBF NN) method for SOC estimation of LiFePO4 battery. BPNN and RBF NN have good characteristics to solve the nonlinear problem. We used discharge and Urban Dynamometer Driving Schedule (UDDS) as training data and testing data. In this research, architecture of BPNN are input layer, one hidden layer with 8 neurons and one output layer. Then architecture of RBF NN are input layer, one hidden layer with 2 neurons and output layer. The experiment used LiFePO4 battery with capacity 2200 mAh, with nominal voltage 4.2 volt. The actual SOC used coloumb counting which are 0 and 1. In this study shows that BPNN and RBF NN can be applied for SOC estimation in LiFePO4 of battery. Both of method have different...

Backpropagation neural network models for LiFePO4 battery

Neural Networks have been used in system control, medicine, pattern recognition and business. The backpropagation neural network (BPNN) appear to be most popular and have been widely used in many applications. BPNN is a supervised learning technique for training multilayer feedforward neural networks. The gradient or steepest descent method is used to train a BPNN by adjusting the weights. The purpose of update numerical weights is minimize error of network between target and output. In this paper, focus with BPNN modeling with data battery for training and testing. We used discharge and Urban Dynamometer Driving Schedule (UDDS) as training data and testing data, respectively Architecture of BPNN consist of input layer, hidden layer and output layer. The otherhand, using BPNN has problem to define amount of hidden neurons. In this study, we used current or voltage as input in input layer, one hidden layer with 8 neurons and one output layer. We used Levenberg-Marquardt algorithm to get fast iteration when...

Fault detection design and simulation based on battery modelling

In electric vehicles, a lithium polymer battery is usually adopted as the main energy storage battery for charging and discharging processes. To manage the battery, Battery Management System (BMS) is needed in the effective and efficient way in order to assure the operation of an EV successful. Failure in the battery means failure in the whole system. In fact, most of the battery damage are found due to the failure of error detection on the battery’s sensors or actuators. Therefore, it requires fault detection on the battery which can work accurately. In this research, the implementation of equivalent circuit models of Lithium Polymer batteries has been done in the Matlab Simulink model to determine the working limit of the voltage and current so that any fault can be detected and isolated, such as overcurrent and overvoltage. The simulation results show that the error detection can be done accurately so that the battery can work optimally.

Current estimation of lithium polymer battery using recursive least square algorithm

Current information on a battery is paramount important in a battery management system (BMS). In many systems, the current sensor is used to get the current value. In spite of its importance, a current sensor is expensive. To overcome this issue, this research aims to design a current estimation algorithm which is based on a current sensorless method where the battery is modeled in a Thevenin equivalent circuit model. The Thevenin model is then formed into autoregressive exogenous (ARX) model and the parameters are extracted by using Recursive Least Square (RLS) algorithm. This research uses lithium polymer battery with a capacity of 2200 mAh and the tests conducted in this research are constant pulse load test and variation load test to learn the performance of the algorithm. The results show that the current estimation has an error of 0.0863A RMSE in pulse load test and 0.6916A RMSE in variation load test.

State of health estimation in lithium polymer battery

BMS is a system to manage the use of the battery and protect it from a condition that led to the failure performance of the battery. One of the components of BMS is State of Health (SOH), which refers to decreased performance of the battery. Estimation methods of SOH are required to reduce the possibility of failure of the battery. This research, systematically designed along through stages of battery test which consisted of the static capacity test, pulse test, and aging cycle test. The results of the testing stages were used to estimate the SOH. Two methods were used to estimate SOH i.e. Coulomb counting and open circuit voltage method. The decreasing of battery SOH could be seen from the changing in the maximum capacity as the number of cycles of the charging-discharging test was performed. The result of SOH estimation indicates that the coulomb counting method was better than the open-circuit voltage method, with the smaller of mean absolute error, mean square error, and mean absolute percent error.

A robust feedback gains for linear systems with multiple delay components

In this paper, a selection of gains to make the systems robust to uncertainty in the delayed plant is proposed. In this work, we remove many assumptions that had been made for so many years in many works. Here the uncertainty can be assumed as natural as possible, i.e., the uncertainty can be treated without assumption of constant bound nor bound as function of the states. It is also shown that using this scheme we can also the state feedback gains can be synthesized. It shown from the numerical results that the method is effective.