“3D TRACKING CAPACITANCE BASED ROS CONTROLLED ROBOT ARM”

Abstract— This paper presents a design procedure of 3d tracking capacitance-based ROS controlled robot arm. A robot operating system (ROS) was modeled to control the robotic arm with 4 degrees of freedom (DOF) to imitate the virtual (Simulation) to reality. ROS is middleware upon which the system can be developed or constructed. The basic controlling operation can be categorized into two sections. The initial process begins with RVIZ simulator control and then control by 3D Cartesian coordinates given by ROS commands to control the arm. The arm covers a quarter of with a radius of 31 cm approximately. Capacitive based 3D tracking mechanism was used to control the robotic arm and update the Simulator by detecting hand gestures. The hand was limited with-in 12 cm cubic .

Keywords– ROS; RVIZ; URDF; Robot Arm; capacitive sensing

Authors- Gunaratnam Pratheepan Auston Institute of Management, Colombo-03, Srilanka, E-mail: pratheepan.guna@gmail.com , Dr. M. K. Jayananda Department of Physics, University of Colombo, Colombo-03, Srilanka, E-mail: kithsiri@phys.cmb.ac.lk

View Full Paper


” DEEP CONVOLUTION NEURAL NETWORK (CNN) USING A MONOCULAR CAMERA AND A 2D LIDAR FOR ROBUST QUADROTOR OBSTACLE AVOIDANCE

Abstract— Quadrotor platforms have become a vital tool in the industry as it is being used as the primary go to technology approach across many sectors including law-enforcement, survey, photography, videography, inspection and etc. However, most users in these industries are not trained or certified pilots, hence introducing safety cautions of crashing drones generated by piloting errors in urban environments causing damage to property and human lives. One of the ways to mitigate such crashes is by introducing robust obstacle avoidance capabilities to the drones in manual control mode. We propose a robust combination technique approach for obstacle avoidance on a quadrotor by a 2D LIDAR and a forward-facing monocular camera using them as a singular system to achieve a higher level of perception about the surrounding environment. 2D LIDAR is used to get an accurate dynamic two-dimensional map of the surrounding environment and the monocular forward-facing camera is used to implement a deep convolution neural network which acts as a classifier network to identify whether the path ahead is safe to continue or blocked. Training data is taken indoors and outdoors for a better generalization for the classifier. We demonstrate experiments showing the capability of this robust obstacle avoidance technique on dynamic obstacles and on static obstacles in both simulated environments (gazebo) and real-world scenarios.

Keywords—Obstacle avoidance; Quadrotor; Deep Learning; Aerial Robotics;

Authors- Yuthika Sagarage, J. K. Jayawardena, Malith Kuruppu Communication Division Arthur. C. Clarke Institute for Modern Technologies Katubedda, Moratuwa, Sri Lanka e-mail: kavindra@accmt.ac.lk

View Full Paper


IDENTIFICATION OF ABNORMAL OBJECTS IN LIVE FEEDS IN THE ATM ENVIRONMENT ”

Abstract— The proposed abnormality identification system is developed to identify and detect abnormalities through a live feed in the banking industry, more specifically in an ATM environment. The research begins with the domain research that is concerned with image processing in terms of identifying the different objects and detecting an abnormal object. The research also covers how the system is to be implemented, the techniques and methods used and how each module is integrated to perform efficiently as a whole. All conclusions are made based on the relevancy to the proposed system and project objectives.

Keywords– Abnormalities, Image, Identification, Live feed

Authors -T.Tharsika Department of Information Technology Sri Lanka Institute of Information Technology, Malabe, Sri Lanka Tharsika.t@sliit.lk , V. R. Dunuwila Department of Computer Systems Engineering Sri Lanka Institute of Information Technology, Malabe, Sri Lanka vandhana.d@sliit.lk , Shashika Lokuliyana Department of Computer Systems Engineering Sri Lanka Institute of Information Technology, Malabe, Sri Lanka shashika.l@sliit.lk

View Full Paper


A REVIEW ON TECHNOLOGY, ARCHITECTURE, APPLICATIONS, CHALLENGES AND FUTURE VISION OF INTERNET OF THINGS

Abstract—The Internet of Things (IoT) is a technology that enables objects to sense changes and respond involuntarily. It facilitates device to human communication, hence giving birth to the ‘Smart’ concept. The overall objective of conducting this review is to understand the present context of IoT. IoT utilizes technologies such as Radio Frequency Identification (RFID), Near Field Communication (NFC), Wireless Sensor Networks (WSNs) and Actuators to identify and track objects while technologies such as ZigBee, 3G, 4G, Bluetooth, and Wi-Fi facilitates communication among devices. The architecture model proposed for IoT is a multi-layered Service Oriented Architecture (SOA) consisting of six layers. IoT applications are already being deployed in various industries such as healthcare, manufacturing, logistics, supply chain and agriculture. The future of IoT inclines towards the introduction of Social Internet of Things (SIoT) and Sensing as a service, which initiated with the advent of cloud computing in IoT. Despite all the positives of IoT, the users also face certain technical, security, privacy, and legal challenges when working with IoT devices.

Keywords-Internet of Things, Sensing as a Service, Smart concept, Service Oriented Architecture, RFID

Authors -V. R. Dunuwila Department of Computer Systems Engineering Sri Lanka Institute of Information Technology, Malabe, Sri Lanka vandhana.d@sliit.lk , T. Tharsika Department of Information Technology Sri Lanka Institute of Information Technology, Malabe, Sri Lanka tharsika.t@sliit.lk , J. A. D. C. A. Jayakody Department of Computer Systems Engineering Sri Lanka Institute of Information Technology, Malabe, Sri Lanka anuradha.j@sliit.lk

View Full Paper


OPTIMUM THRESHOLDS FOR EFFECTIVE MARKET SURVEILLANCE VIA NOVEL PRICE MODELS

Abstract— Unusual movements in prices may indicate manipulative behavior or be the result of a market anomaly that could serve as a precursor to a sequence of events that could lead to major losses. Detecting these events in order to ensure investor confidence and an efficient market is the province of market surveillance. This paper presents novel methods by which to model prices and place limits on its movement ensuring that such unusual behavior is detected in time to effect preventive measures. Market surveillance is often affected by a series of software configurable alerts that depend on a set of thresholds to detect when prices are outside certain desired limits. An optimum set of such thresholds can be estimated by analyzing historical prices to estimate the likelihood of observing unusual movements leading to a more effective solution. In one contribution we estimate the random process governing the movement in the price and simulate a series of realizations drawn from the same underlying process to gain insights into unusual behavior that could manifest. In this modeling, we estimate the distribution of the sequence of random variables governing the prices and generate a series of realizations or paths with the same underlying distribution. We also develop an equation to predict the expected deviation in price between two points in a sequence of prices and a measure of the uncertainty in this deviation. In another contribution we model prices as a linear regression on consecutive prices to estimate its movement and arrive at an estimate for the distribution in the error of such a prediction. By these techniques we estimate the trend and maximum deviation in price that could be expected over a sequence of prices in order to optimize the alert thresholds. Through these analyses we also observe that the variance in the price is dependent on the number of samples in a sequence of prices over which the measurement is made due to the behavior of the random process governing its movement and propose that the variance be estimated over a fixed number of consecutive prices ensuring a more stable and consistent estimate. By these contributions we envisage a more stable trading environment with lower risks facilitated by an effective surveillance solution enabled by a set of optimum thresholds that are estimated by a more accurate characterization of the movement in the price.

Keywords— linear regression; random walk; volatility; market surveillance;

Author – Asoka Korale, Ph.D. London Stock Exchange Technology, Malabe, Sri Lanka, asoka.j.m.korale@gmail.com

View Full Paper


IMPLEMENTATION OF A ROBUST GROUND-STATION FOR LOW EARTH ORBIT (LEO) SATELLITE COMMUNICATION

Abstract—The Ground station developed by ACCIMT for receiving Amateur band signals utilizes a central Terminal Node Controller (TNC) based on a Gaussian Minimum Shift Keying (GMSK) modem. The TNC is responsible for modulation and demodulation of the communication data. A commercially available actuator and a controller were used for antenna pointing. A computer interface for in-house developed hardware system which is compatible with RS-232 protocol was developed. A Graphical User Interface (GUI) software application running on a standard computer acts as the satellite’s operating environment. A commercially available transceiver (ICOM 9100) was used to transmit telemetry control to the satellite and receive the satellite Continuous Wave (CW) beacon and data. In-house developed software was used for satellite tracking and antenna rotor controlling Two Line Element (TLE) file of an interesting satellite is obtained from available sources to feed into the said software. The software generates a serial data stream to the rotor controller via the protocol converter, which converts RS232 data to controller input voltages. Three open-source JavaScript (Google maps API, TLEjs and jspredict) libraries were used as Application Program Interface (APIs) to develop the said software with electron js. These APIs get the TLE as an input and output the accurate orientation data, which is then fed into the antenna rotator so that it can maintain the antenna in the direction of satellites.

Keywords—Satellite, Ground Station; Tracking; LEO; TLE

Authors- R.A.D.K. Sampath, J.K. Jayawardena, Y.G.C. Jayalath, S. Vigneshan Communication Division, Arthur C Clarke Institute for Modern Technologies Katubedda, Moratuwa, Sri Lanka e-mail: radkaveen@gmail.com

View Full Paper


PILOT RESEARCH PROJECT ON REAL-TIME MONITORING OF RIVER WATER LEVEL AND IRRIGATION MANAGEMENT AT GRAVITY OUTLETS (NILWALA RIVER)

Abstract—Flooding has become a major problem in most of the countries around the world, causing huge damages. It affects the farming community, incurring disturbance to day to day activities, mainly by damaging the farming land and infrastructure. Although there are methods to forecast rainfall or to track storm path precisely from the satellite images in the world, the need for real-time monitored data such as river water level is essential to make a reasonable decision on the actions necessary to prevent flooding in an automated manner. The rivers are usually located in the low lying lands and agriculture productions usually utilize river water in these areas. In wet zone areas due to heavy rain small streams and canals of the main river which are feeding agriculture land gets more water compared to upstream sections. In this situation, Gravity Outlets (GO) come into action where the decision is taken to transfer the water by using gravity or forced pumping, after comparing the upstream water level (riverside) and downstream water level (Yaya or protection side). This system is widely used in Europe. At present, all the pump houses are fixed with river and landside conventional manual gauges to monitor the water level and also to make decisions for pump operation. Since, this is a manual, individual dependent and less reliable system, GO operation is not efficient. It is observed by the Department of Irrigation that one-centimeter level gravity difference can save pumping costs in the order of magnitude. Last few years it has faced more disputes due to less finance allocation from the government for these routine tasks. Paddy fields of farmers and low land residential areas are severely damaged during the flood instances. Through an automated, reliable and quick data providing solution based on the Internet utilizing GPRS (General Packet Radio Switch), information can be delivered to the action points that can streamline the operation of gravity outlets. It will greatly reduce the cost of daily operation and farmer conflict for the Department of Irrigation.

Keywords—Irrigation; SCADA; Flood; GO; GPRS; River; Pumping; Internet

Authors – P. D. S. Pushpakumara, J. K. Jayawardena, R. P. Dasanayaka, R. A. D. K. Sampath, D. Pieris, G. D. N. de Silva Communication Division Arthur. C. Clarke Institute for Modern Technologies Katubedda, Moratuwa, Sri Lanka email: kavindra@accmt.ac.lk

View Full Paper


USE OF UNMANNED AERIAL VEHICLE (UAV) PLATFORM FOR 3D MAPPING AND NAVIGATION OF SUBTERRANEAN LANDSCAPE USING 3D LIDAR SENSOR

Abstract— This paper presents the formalization of the hardware and software implementation for an Unmanned Aerial Vehicle which will be used for 3D mapping and real-time 2D obstacle avoidance using a 3D Light Detection and Ranging sensor. The proposed implementation can be used to map and navigate in areas with low visibility, limited or no Global Navigation Satellite System coverage and without any prior information about the site such as a in a subterranean landscape. The navigation is based on a Vector Polar Histogram Plus, but the “construction of cost function” is modified to navigate through unknown environments. This also means that the system is fully autonomous. The 3D mapping is based on “Velodyne SLAM”, which simultaneously localizes and maps its environment in 3D.

Keywords— UAV, LIDAR, Subterranean Navigation, 3D Mapping

Authors – Malith Kuruppu, Yuthika Sagarage, J. K. Jayawardena Communication Division Arthur. C. Clarke Institute for Modern Technologies Katubedda, Moratuwa, Sri Lanka kavindra@accmt.ac.lk

View Full Paper


ASSESSMENT OF TURBIDITY OF WATER USING IMAGE PROCESSING TECHNIQUES

Abstract— The research base is to prove the novel concept and a simpler, to measure turbidity is by taking pictures of water samples with a camera at a constant lumen level, and using to identify the level of turbidity of the taken sample based on image classification algorithm. It is observed in Sri Lanka that there are generally expensive turbidity sensors, therefore, lack of automated water regulation methods, which has proved to be ineffective. Therefore, an online turbidity sensor is an important measurement device which is sought by water management authorities across the local and international market, especially one of which is potentially inexpensive and ease of use. The National Water Supply and Drainage Board (NWSDB) of Sri Lanka expressed their interest in such an efficient and low-cost device for better water treatment. Also, samples with known turbidity levels for verification purposes supplied by the Biyagama water treatment plant.

Keywords-Turbidity; Water; Image Processing, image classification

Authors -Oshadi Thalpawila, Yuthika Sagarage, J. K. Jayawardena, R. A. D. K. Sampath, Thilina Tharanga, Malith Kuruppu Communication Division Arthur. C. Clarke Institute for Modern Technologies Katubedda, Moratuwa, Sri Lanka kavindra@accmt.ac.lk

View Full Paper


HYBRID ENERGY HARVESTING AND STORAGE FOR WIRELESS SENSOR NETWORK

Abstract- Advanced development of wireless sensor network is comprised in the expansion of the research acquisitions. In case of integration, self-energy harvesting and long-lasting energy management have become challengeable topics in the development of wireless senor networks (WSN). In addition, continuous dataflow is a vital requirement of a successful product by means of WSN. Therefore, this research paper outlines a system that can overcome drawbacks in existing studies, and it introduces new features to WSN deployment. The proposed system introduces a wireless sensor node which consists of three components; Hybrid self energy harvesting system, Hybrid energy storage and uninterrupted wireless data transmission between nodes and the end user. Solar and wind energies are used as harvesting bases to fulfill the continuous power requirement of the sensor node. The hybrid energy storage assists to store the energy in the middle of the power flow from the harvester to the node. In an instance, Lithiumion batteries and Supercapacitors are combined as energy storage resources. Moreover, the system supports to analyze and transfer the health status of each key element of power management in both energy harvester and storage. The WSN is developed by constructing a direct communication link to the remote monitoring server instead of an Internet connection. That ensures the hundred percent real time data delivery to the end user without any data losses. Additionally, a special node failure detection is included. The consolidation of the three components executes as a smart system that can be used to aggregate required sensor data and transfer them to the end user. The prototype is decided to test in a demonstrative over the bridge and uses the accelerometer as the sensor to collect risk predictable data from it.

Keywords – WSN, Hybrid energy storage, Hybrid self-energy harvesting, RF Protocol

Authors -Anuradha Jayakody1 , Panchalee Illeperumaarachchi2 , Dasun Chathuranga3 , Anjelo Rajapaksha4 1-4Sri Lanka Institute of Information Technology (SLIIT), Department of CSE – Sri Lanka, Malabe 1anuradha.j@sliit.lk, 2 panchalee.i@sliit.lk, 3 it14134036@my.sliit.lk, 4 it15033260@my.sliit.lk,

View Full Paper


BIOMETRIC SMART SECURITY SYSTEM WITH CHILD CARE FOR A SMART SOCIETY

Abstract— This research is mainly focused on Infant movement detection and alerting, in order to enhance their security within the home premises. As the first move, the research focuses on the identification of the human and classifying whether an adult or a baby. Then a model was built up in three classifications to identify static and dynamic positions of the infant, through Image Processing and analysis. In order to enhance the accuracy of the custom classifiers an already trained model using 1 million image set was retrained by customized image sets. To present this research as a smart home solution modern technology were used in implementing the close connection between the infant and the parent.

Keywords— Custom classifiers; Frame analysis; Image Processing; Object detection

Authors -Shashika Lokuliyana1 , I.U.Mundigala2 ,G.H.Ashan Sanjeewa3 Department of Computer Systems and Networking,Department of Information Systems Engineering Sri Lanka Institute of Information Technology Malabe,Sri Lanka 1 shashika.l@sliit.lk, 2 iyanthiupeksha@gmail.com,3 ashansanjee@gmail.com

View Full Paper


EATND – IOT BASED SMART ATTENDANCE SYSTEM FOR A SMART CAMPUS

Abstract— Internet of Things (IoT) is the driven concept behind incorporating a large number of different and heterogeneous systems. With the implementation of smaller sensors and devices, a large variety of devices can be connected using link-layer technologies to provide various types of services. In this paper, authors focus on implementing a Smart Campus environment by implementing an IoT platform combining Wireless Sensor Network (WSN) technologies for communication and localization. Added-value services such as administration, information distribution, student localization along with a mobile application. The system aims at providing administration services to the campus authorities as well as providing information services to the students. Furthermore, the paper presents the implementation and the evaluation of the system performed with the collaboration with campus and its students.

Keywords— IoT; WSN; RFID; Smart Campus

Authors -Shashika Lokuliyana1 , Anuradha Jayakody2 , G.R. Samaraweera3 , W.M.T.M. Bandara4 , W.V.D.S.A. Wevita5 , P.K.B.M.H. Bandara6 Department of Information Systems Engineering, Faculty of Computing Sri Lanka Institute of Information technology, Malabe, Sri Lanka 1 shashika.l@sliit.lk, 2 anuradha.j@sliit.lk, 3geeth.sam1@gmail.com, 4bandaratharindu07@gmail.com, 5 shehan19949@gmail.com, 6hasarangabandara123@gmail.com

View Full Paper


AN INTELLIGENT TRAFFIC SIGN RECOGNITION SYSTEM FOR SRI LANKAN DRIVERS

Abstract—Automatic traffic sign detection and recognition system plays a crucial role in enhancing the safety of road users and of motor vehicles. It provides accurate and timely warning alerts of driving conditions to drivers with a minimal human effort. The effectiveness of such system depends on many factors such as vehicle speed, background complexity, lightning and weather conditions. In this paper, we propose an automatic traffic sign detection and recognition system suitable for Sri Lankan road system. The proposed system consists of two stages, traffic sign detection stage and traffic sign recognition stage. We adopt a Haar Cascade classifier for traffic sign detection and Convolutional Neural Network (CNN) approach for traffic sign recognition. Experimental results demonstrate the effectiveness of the proposed automatic traffic sign detection and recognition system for different vehicle driving speeds

Keywords— automatic traffic sign detection; recognition; haar cascade; convolutional neural network

Authors – I.B Hewanayake 1, A.K.U Ranasinghe 2, A.M.A.L Gunarathne3, C.K.W Seneviratne 4, W.N.B.A Geeth Priyankara 5

View Full Paper


A Regulatory framework for a Smart Hydroponics System

Abstract—Climate change is instigating extreme weather events, which include heat waves, droughts, strong winds and heavy rainfall, and these events can impact the availability of water for agriculture. The unpredictable droughts and floods in Sri Lanka have made the production of food crops difficult. Excess water and decreased soil moisture have resulted in a sharp decrease in crop yield. Thus, there has emerged an urgent need to employ new agricultural technics that use less water and nutrients, and which are not easily affected by external environmental variables. Aeroponics is fast becoming an opposite solution to the aforementioned challenges posed by climate change to agriculture. Furthermore, this method of growing crops can be used to grow crops in the space and other planets. However, the technic of growing plants using aeroponics is mechanically convoluted, is vulnerable to malfunction and requires accurate regulation and control of water and nutrients. Ergo, the focus of this research is on constructing an automated system that regulates and monitors the aeroponic system to reduce to a greater extent the challenges this agricultural technic faces in becoming a commercially viable solution to the modern agricultural problems. This will be accomplished by the use of ultrasonic atomizers, pH and EC regulators, gas sensors, temperature regulators, humidity regulators, and a single-board computer, while a mobile app and a web interface will be employed to allow the user to monitor and interact with the system. The user will also be warned in case a malfunction takes place in order to control the amount of damage to the plant.

Keywords— Vertical farming, Hydroponics, Aeroponics, Greenhouse systems, Internet of Things

Authors – Anuradha Jayakody1,Shashika Lokuliyana2,P. D. S. Silva3,V. R. Dunuwila4

View Full Paper