2019-2020
Main block
Project Title: Application of Artificial Intelligence In The Detection and Diagnosis of Neurological Disorders
Group Member: Bernard Atiemo Asare & Sabrina Awuni Lamie.
Supervisor: Dr. Godfrey A. Mills
Co-Supervisor: Dr Elsie Effah Kaufmann
Abstract:
Neurological Disorders are diseases that are associated with the central and peripheral nervous system of humans. Some of the most commonly encountered neurological disorders include Parkinson’s disease, stroke, Alzheimer’s disease, epilepsy, migraine and headache disorders among others. These disorders severely affect the health of an individual and can lead to effects such as partial or complete paralysis, muscle weakness, headaches and even death. Early detection of potential neurological disorder through computer-aided screening processes may lead t quick intervention measures to avid major health challenges. This thesis provides an artificial intelligence-driven software solution that could be used on both smart mobile devices and computers by physicians for effective screening of patients for the detection and classification of neurological disorders.
An artificial neural network was designed and trained using acquired features extracted from the patient’s physiological information from the screening. Three common neurological disorders prevalent in Ghana were used for the study and these are Parkinson’s disease, epilepsy and stroke. The neural network had an input layer with 31 neurons, one hidden layer with 25 neurons and an output layer with 4 neurons. The network was trained using a patient physiological dataset. Out of the dataset, 70% was used for training and 30% was used for testing. The capability of a neural network to diagnose and detect neurological disorders was tested using numerically simulated data as well as real screening data from patients. Blind tests were also carried out by neurologists at the Korle-Bu Teaching Hospital. Results from both the simulation and the experiment reveal that the artificial intelligence model could accurately detect and predict conditions of neurological disorder and classify the type of disorder with a detection accuracy of 100% for Parkinson’s disease, 100% for epilepsy and 100% for stroke.
Through this artificial-intelligence driven solution, the tasks of neurological experts or related specialists could be shifted down to primary care physicians to undertake the screening and detection of neurological disorders and refer the critical conditions to the experts for attention. This is a much needed solution for Ghana and other developing countries with a limited number of neurological experts and resources. It may also be a useful resource for training medical and health-related students on neurological disorder detection.
Project Title: Smart Shuttle System For University Of Ghana Campus
Group Member: Henry Osei Agyemang, Alfred Domegil Degbang & Elvis Bosomafi.
Supervisor: Dr. Robert A. Sowah
Abstract:
A smart shuttle system is a system that incorporates several technologies such as real-time tracking of vehicles as well as other services such as providing the estimated time of arrival of the shuttle to a particular location and a means for users to pay for their travels. In this work, a smart shuttle system was designed and implemented to help the University of Ghana students to locate, in real-time, the position and estimate the arrival time of shuttles via a web application and a mobile application. The system will also include a payment system to provide a means for students to pay for their rides without physical cash. This project comprises an integration of hardware and software. A real-time Google map and Arduino based vehicle tracking system is implemented with the Global Positioning System (GPS) and Global System for Mobile Communication (GSM). An Arduino UNO is used to control the GPS module to get the geographic coordinates at regular tie intervals. Then the GSM module transmits the location of the shuttle the web or mobile application that will be developed in terms of latitude and longitude. Finally, the web or mobile application uses Google Maps to show the location of the shuttle in real-time hence making it possible for students to monitor a moving shuttle using their smartphones or laptop. The mobile app has a feature that allows students to pay for a ride in the shuttle through its payment system. The hardware tracker is placed in a shuttle and it obtains the location of the bus and transmits the location via an Application Programming Interface (API) endpoint to the backend of the web application which sends the data to the firebase for storage. When a user signs in or signs up on the web application he or she is redirected to the map page where he or she can see the location of the bus on Google Maps as well as the estimated time. When the user wants to pay for a ride, there is a payment page on the web application where he or she can pay with either mobile money or credit card. Our system was tested for the various functionalities mentioned above to make sure that the system was fully operational. As such, the primary objective of this project which is to allow students to be able to locate the position of the shuttle on campus as well as the provision of a payment system for students to be able to pay for using shuttle services was achieved. The main recommendation for this work is the addition of more features to make it more accessible for people to be able to use it efficiently.
Project Title: Development of a Fraud Alert System for Mobile Money Subscribers
Group Members: Isaac Adzah Sai & Richard Osei Sakyi
Supervisor: Dr. Wiafe Owusu-Banahene
Abstract:
Mobile Money Fraud started in the year 2011. Fraudsters have constantly been finding and using the weakness of the platform to trick people into falling for their schemes. The number of mobile money users has already exceeded the bank transactions in Africa.
It is having a massive effect on Agriculture, Banking, Insurance, Healthcare, and other sectors in Africa. The platform hasn’t been reported to have been hacked, but fraudsters are not tracked and prosecuted. For now, only the number or mobile device used to commit the fraud is blocked. This gives the fraudster the privilege to register and use a different number or mobile device since there is no one checking if the person registering with a valid identification (ID) card has committed fraud with another number before.
In this work, two platforms are created, one to help automate the registration of mobile money users for telecommunication networks by using Artificial Intelligence to speed up the registration processes. To Speed up the registration, Optical Character Recognition (OCR) is used to extract relevant information for registration from an image of the user’s ID card. The image can be obtained by using a computer’s webcam or by uploading it from the computer. The information is then rectified by the administrator then stored in a database. The data is the linked to an individual helping to track all multiple numbers he/she owns. There will be constant updates and syncing of data to keep it valid at all times.
The second platform, a mobile application was built and made available to subscribers. This is used as a secure platform for mobile money transactions. Users can file a complaint or report a number that has attempted or succeeded in defrauding them. The reported numbers are sent to the telecommunication companies quickly for investigation and when the investigation shows that the reported number has been used in a fraud, it is blocked and blacklisted. The mobile application has a broadcast receiver which listens to incoming calls or messages no subscribers' phones. The ID of the number making the incoming call is cross-checked with the caller ID that has been stored in the database of the telecommunication companies. When a fraudster is detected, the subscriber will be alerted that the incoming call can be coming from a fraudster. He or she can then allow the application to permanently block that caller.
Project Title: Quality of Service Monitoring System for GSM
Group Members: Rockson Bernard Asare, Elvis Ukoji Solomon & Daniel Albert Ludwig
Supervisor: Dr. Wiafe Owusu-Banahene
Abstract:
The quality of service is accessed in terms of key performance indicators based on statistics generated from drive tests or network management systems. The technique consists of employing an automobile containing mobile radio network air interface measurement equipment which will detect and record a good sort of the physical and virtual parameters of mobile cellular service during a given geographic area.
By measuring what a GSM subscriber would experience in any specific area, GSM operators can make direct changes to their networks that provide better coverage and repair to their customers.
This drive test is not done every day. We found a problem with moving from one geographical area to another to measure the service quality. Our solution to this problem is to develop a quality of service monitoring system to measure and monitor network quality.
With the Mobile and Web App, the user can monitor the service quality themselves and get to choose the better network in their area. It will also help telecommunication companies to also monitor their networks in specific geographical areas in the comfort of their workplace.
Project Title: Smart Agricultural Monitoring and Automation System for Farms
Group Members: Sagoe Frederick Hazel & Aaron Kweku Essuman
Supervisor: Dr. Wiafe Owusu-Banahene
Abstract:
Agriculture has been practiced in nearly every country for ages. Agriculture was the key development in the rise of sedentary human civilization. Agriculture has been done manually for ages. Agricultural practices stand to benefit from the advancements in technology leading to smart agriculture. Internet of Things (IoT), one of the emerging trends in technology, can play a very important role in creating smart agriculture. Our work aimed at utilizing IoT technology, web software, mobile software, and Artificial Intelligence (AI) to create an integrated monitoring and automation system for farms. We developed an IoT system powered by Arduino with a Temperature sensor, Moisture sensor, water level sensor, DC motor, and a GPRS module. This IoT based Agriculture monitoring system makes use of wireless sensor networks that collects data from different sensors deployed at various nodes on a farm and sends it through the wireless protocol to web and mobile software system. When the IoT based monitoring system starts it checks the water level, humidity, and moisture level. It sends an SMS alert onto a mobile phone and web application about these levels. When the sensors sense that the level of water has gone down, it automatically starts a water pump to fill the water trough and a red Light Emitting Diode (LED) glows showing the water pump is on. If the temperature goes above the level, the fan starts. These are all displayed on the Liquid Crystal Display (LCD) module. The web and mobile application at as well as the LCD show Humidity, Moisture, and water level with date and time, based on per-minute readings. Temperature can be set on a particular level, depending on the type of crops being cultivated. It is possible to close the water forcefully. There is a button on the IoT from which the water pump can be stopped. The system has an AI prediction model to make approximations of possible temperature outcomes based on the recorded values collected from the greenhouse.
Project Title: Drone For Supporting Emergency and Medical Services
Group Members: Kwadwo Osei Junior, Peter Turkson Asamoah & Richard S. Nsiah-Agyemang Jnr.
Supervisor: Dr. Godfrey A. Mills
Co-Supervisor: Dr. Elsie Effah Kaufmann
Abstract:
Drone or UAV is an aircraft which is not used to carry passengers and can be made autonomous or controlled remotely by a pilot. Drones are classified based on the altitude range, endurance and weight, and support a wide range of applications including military and commercial applications. Drones can be used for various applications in the health sector, military or for photography. In this project the problem of medical supply delivery to medical emergency centers and disaster sites is solved using autopilot methodology and a design that can meet the weight specifications of the medical supply.
The goal of this project is to design and develop an automatous drone using a quad-copter configuration and equipped with communication facilities to deliver emergency medical supplies and other services to rural health centres or disaster centres. The communication module is to facilitate instructional guidelines to the health professional in the event of the need for technical support. The quad-copter was designed to meet a payload weight of 1 kg and coverage distance of 1 km. The drone was equipped with GPS navigation feature and programmed to facilitate auto navigation from the drone control centre to the delivery destination. The design was implemented in a simulation environment and a prototype of the drone was also developed and tested for functionality. Our project work yielded a successful development of quad-copter with autonomous navigation. Real-time communication between by-standers and emergency service personnel at emergency site and the professional doctors back at the distribution Centre was also successfully implemented. The Drone can be applied in local hospitals and medical centres to deliver medical supplies to the people who need immediate medical assistance.
This report contains an in-depth guide to the hardware and assembly of a drone system, the software required for a basic drone set-up as well as additional software needed or autonomous operations, and finally the results of the project along with the difficulties and setbacks that were encountered.
Project Title: Computer Vision Based Driver Fatigue Detection and Alert System
Group Members: Kevin Amexo & Kwadwo Adu Boakye-Yiadom
Supervisor: Dr. Wiafe Owusu-Banahene
Abstract:
Vehicle accidents are a common occurrence all over the world, large portions of which are fatigue-related. There is therefore the need to come up with solutions to the high number of vehicle accidents caused by fatigue. The system developed comprises a microcontroller, a camera that captures the driver’s face and a speaker. The microcontroller receives a video stream from the camera, and analyses the eyes and mouth of the driver to detect signs of fatigue. This was accomplished using Haar Cascades. The system developed is able to detect fatigue with a high accuracy. The Computer Vision Driver Fatigue Detection and Alert System therefore provides a solution to the problem of fatigue related vehicle accidents. This project could be furthered by integrating the system with self-driving cars, to automatically switch into autopilot when the driver continuously exhibits signs of fatigue.
Project Title: Solid Waste Monitoring and Revenue Management System
Group Members: Amma Frimpong-Boateng & Annie Asabea Boadu
Supervisor: Dr. Wiafe Owusu-Banahene
Abstract:
Effective and efficient waste management is a very important aspect of our every society. Due to increasing population and advancement in technology, there’s a drastic increase in the amount of waste being produced in Ghana and the world as a whole. The main sources of waste are domestic, commercial, industrial, municipal and agricultural wastes. If waste produced is not properly managed it can have many negative effects on our environment and the human resources of the country. Several companies have systems in place to manage waste in our society. These existing systems require improvement in order to monitor the waste as it is being stored at the place of generation. Moreover, revenue collection and management is still a challenge for waste management companies: in most cases agents go round to manually collect revenue. In order to offer solution to these challenges, we developed an integrated hardware and software system referred to as, Amanie. Amanie seeks to offer complete monitoring and revenue management. We designed a bin which can monitor waste being generated using ultra sonic sensors to measure waste levels and automatically send information captured to a server via a wireless mobile network connection. This information can be accessed by administrators via a web application, so that waste can be collected on time. Amanie also features an android mobile application that allows users and collectors to access information of waste levels in bins and also includes a revenue system that allows users to make payments via a mobile money. All activities and information generated from the mobile application is then passed to a database server which can be accessed by the administrator via web application. Our system will improve waste monitoring and revenue management since most of the manual processes have been automated.
Keywords – Waste, bin, collector
Project Title: Drone Flight Controller System For Land Mapping and Survey
Group Members: Priscilla Owusu-Prempeh & Gerald Dugbatey
Supervisor: Dr. Nii Longdon Sowah
Abstract:
Drone is classified as an aerial vehicle that does not carry a human operator, uses aerodynamic forces to provide vehicle lifts, can operate independently or remotely, can be extended or recoverable, and can carry a lethal or non-lethal payload. It is operated either individually by on-board computers or by the pilot's remote control on the ground. The goal of the Project is to develop and implement the drone for land surveying and mapping applications, based on the open source Autopilot called Ardupilot framework. Drone flight controller system is simulated and built for land mapping purposes. The Drone control system is developed and simulated in MATLAB / Simulink. The simulation shows the built Drone being run and managed very stably. Microcontroller-based drone control system is also developed. In this case, the Drone used an RC transmitter and receiver for remote operation.
Evidence from the literature review and research undertaken indicates that the use of the drones makes it possible to cover a wide variety of terrestrial and aerial approaches and that it can supplement or complement other means of surveying and data collection. The usage offers the ability to get close to the target without being stuck on the land, as well as benefits from the operating environment, as risky, challenging environments can be reached from a distance. Data can be obtained quicker, faster, easier and more often. Time savings exist at the calculation level, but more time is required for the post-processing of the data relative to terrestrial approaches.
Project Title: Real-Time Face Recognition and RFID Attendance System
Group Members: Kwaku Adusei Okyere & Francis Kofi Anane Wormenor
Supervisor: Dr. Nii Longdon Sowah
Abstract:
Regular attendance is key to giving students the chance and conducive environment to gain knowledge, any failure to attend class interferes with the learning process. While research affirms the significance of teacher effectiveness on student academic performance, even the most efficient teachers cannot impact learning if students are not present in class physically or remotely.
As the importance of attendance have been stated, the means of recording or tracking of attending data is also a key issue. Attendance records are traditionally taken by passing a sheet of paper in class or mentioning student names and marking responses in a register. This can take too much time and effort depending on class size, moreover, the reliability of the papers or registers is questionable.
Face recognition is gradually becoming a very prolific and efficient way of identification nowadays. It is being adopted by many top companies like Apple, Samsung, Microsoft for identification and verification of users. We chose to utilize face recognition as a key tool in creating a more efficient, faster, and highly autonomous attendance record system. Also, by using RFID technology as a supporting tool, we aim to reduce the deficits of face recognition to a minimum.
By coupling face recognition and RFID technology we were able to create a better attendance management system that is not only faster and more efficient but also generates and stores electronic records which is more reliable and easier to perform analysis on.
On implementation and testing, we deduced that have more images per person as the dataset, preprocessing images (realigning faces with dlib) and tweaking parameters improves positive recognition rates. Also, some factors like lightning, eyewear, etc. affect both detection and recognition. Lastly, though our project used one camera for demonstration, we believe having multiple cameras mounted at vantage points would also facilitate better performance of the face recognition system.
Project Title: IoT-Based Automatic Vehicle Detection System
Group Member: Abdul-Shahid Mohammed
Supervisor: Dr. Robert A. Sowah
Abstract:
The high demand for automobiles has also increased the traffic dangers and road accidents thereby putting people's lives under high risk. The period between the occurrence of the accident and the arrival of emergency services to the scene is a significant antecedent of survival rates after diagnosis of the accident. Thus, delay in getting the ambulance to the site of the accident and the traffic congestion between the location of the accident and the hospital increases the victim's risk of death. The automatic accident detection system comes to the rescue to resolve the problem. With the emergence of the internet and computational era, it is highly desirable to have a smart monitoring and reliable system in vehicles to effectively relay information when accidents occur, this can be achieved through the introduction of the Internet of Things (IoT) technology. This project comprises an integration of hardware and software. A real-time Google map and IoT based accident tracking system are implemented with the accelerometer sensor, WiFi module, Global Positioning System (GPS), Global System for Mobile Communication (GSM) and a mobile app. A microcontroller is used to control the onboard sensors and continuously monitor the system. In a situation where an accident occurs, the accelerometer sensor sends a signal to the microcontroller in the system, it then sends an alert message which includes the location and user information to the rescue team via a mobile app alert. Then the GSM module transmits the location of the accident that will be developed in terms of latitude and longitude. Finally, mobile application uses Google Maps to show the location of the accident in real-time hence making it possible for emergency personnel to track using their smartphones. The hardware tracker is placed in a vehicle and it obtains the location of the vehicle when an accident is detected and transmits the location via an Application Programming Interface (API) endpoint to the backend of the mobile application which sends the data to the firebase for storage. When a user signs up on the app, the necessary information required by the emergency team is stored in the database for retrieval when an alert is created to aid the team. The system was tested for the various functionalities mentioned above to make sure that the system was fully operational. As such, the primary objective of this project which is to allow emergency personnel to be able to locate the position of vehicle accidents as well as provide them with the details of the user involved in the accident. This project seeks to detect accidents in significantly less time and transfer the fundamental information to the nearest health covering the geographical coordinates and the time where the vehicle met the accident. As there is scope for enhancement and as the main recommendation for this work, the addition of more features such as a wireless webcam to capture the images which will help provide driver support may be incorporated to improve efficiency.
Project Title: Wearable for the Medically Fragile
Group Members: Augustine Yeboah-Afari & Reginald Darko Asiedu
Supervisor: Mr. Stephen Kanga Armoo
Abstract:
Smartwatch is a wearable minicomputer or a mini smartphone in the form of a wristwatch. Smartwatches provide a touch screen interface; this smartwatch has many potential capabilities, like message notifications, GPS navigation, and calendar synchronization, and of course, a Bluetooth connection to your phone which helps you to call or send and receive messages. It also acts as a fitness tracker that can count your footsteps, measure the distance covered, calories, monitors your heart, pulse rate, tracks your sleep and even some smartwatches calculate other important metrics that you might need. Many Senior Citizens are retired and most of them suffer from major and minor health issues that affect their routine life. To take care of the seniors suffering from such diseases it is necessary to track their health status by regularly checking their heart rate, temperature, etc. Our proposed system serves as a solution to this problem; it keeps tracking the health status of the seniors and sends the health status like heart rate, humidity through SMS to the respective caretaker once in every min. Also, if the elder person falls or collapses down it will send an immediate SMS and location to the caretaker. This system is powered by Atmega328. It consists of an LCD display, GPS, temperature sensor, fall detection sensor, and heart rate sensor.
Project Title: Smart Intrusion Detection Alert and Alarm System
Group Members: Kpeglo Emmanuel & Tweneboanah Richmond Maunge
Supervisor: Dr. Nii Longdon Sowah
Abstract:
This report contains a proposed system that will ensure that the welfare of the facilities of users are ensured. Smart intrusion detection, alarm and alert system detects intruders in a facility or home and sets off an alarm while alerting the owners of the system. To ensure that no false alarm is set off, a facial recognition algorithm is implemented in the system. This facial recognition is achieved using computer vision and machine learning.
Robberies are common problems in places where security systems are unavailable. Facilities and homeowners who have been attacked before, or are victim to such attacks know how terrible it feels, to discover that someone has broken into your home and stolen your money or properties. It is important to make people know that such problems can be reduced or be prevented by installing security systems at their homes or facilities, to protect their properties and also prevent potential break-ins by burglars; hence the need for this project.
Designing a smart intrusion detection system that can set off an alarm and also alert the owners of homes remotely, can solve the problem stated above. This system uses a facial recognition system to prevent false alarm. Most systems set off false alarms and this can be a problem. The system also has a Passive Infrared (PIR) sensor to detect the presence of intruders.
There are lots of security systems on the market of which most aren’t affordable by the average individual. Some are easy to install and some are quite the opposite. Some of the systems are less likely to prevent false alarm since what they mostly do is detect motion and set off an alarm. The system we have designed is easy to install and very affordable.
The system protects valuables, allows remote access to the system, notifies if there’s an intrusion and it also helps you keep your mind at peace. The system can be used in residential, commercial, industrial and military properties for protection against intruders.
The smart intrusion detection, alarm and alert system is efficient, affordable and easy to install. The system can be further improved for home automation and also used as a potential fire detector or alarm.
Project Title: IoT Based Intelligent Gas Leakge Detector Using Arduino
Group Member: Leonard Nii Tettey Nyaban.
Supervisor: Mr. Stephen Kanga Armoo
Abstract:
In this modern age where technology is at the forefront of every commercial and domestic sector and endeavor, technology has not only led to increased ease in performance of tasks and labor but also to a high level of safety, security and refuge in various aspects of life. In Ghana, the occurrence of Liquefied Petroleum Gas fires and explosions is an all too common occurrence which has plagued the country greatly. With four major gas leak explosions occurring four times within a span of ten (10) years, a statistic that is considered high on the National Petroleum Authority’s (NPA) scale. Regardless of this the country does not have any automatically implemented means to prevent the occurrence of these disasters with all the proposed methods relying on detection and warning by a human. With the continually increasing use of Liquefied Petroleum Gas in the country, it is necessary for safety measures to be put in place to reduce or prevent these disasters and save life and property. An automated gas leakage detector and alert system is needed to detect Liquefied Petroleum leaks and provide immediate warning for security measures to be taken to prevent loss of life and property.
Project Title: Farm Management System With Machine Learning
Group Member: Parry Bernard Nana Gyimah, Archibald Kwadwo Boateng & Nana Banyin Tandoh
Supervisor: Dr. Robert A. Sowah
Abstract:
Farm Management is simply the science of optimizing the use of resources in the farm component of farm-households, Farm management systems give farmers an instant overview of the day to day weather forecast or yearly history for every field. Based on the farmers, farm management systems may differ from one farm to another, but the significant concern of farm management is the type of farm. The system is designed for the particular need of the farmers to carry out operations smoothly and effectively. In our system, there is a drone and a drone camera for surveillance of farmland, a web server which stores, processes and delivers web pages to the users, a database server to store data from farmland and system and a weather API to allow access to current weather data on farmland. The drone camera after monitoring and obtaining live feeds from the system transmits the sensor data and video stream or image capture to the web socket server. Implemented in the system is a Mobile App written in Java and runs on the Android Platform and it provides a live camera feed of what the drone sees when mounted on the drone Platform. The quality of the video stream is highly dependent on the Camera Resolution of the Host device and also the quality of the internet connection, however since a high priority is given to the smoothness of the feed, the system may intentionally degrade the quality of the video feed to maintain the frame rate. It also provides you with the option of deploying your trained ML models to the cloud for inference and testing. The web portal offers an easy to use and intuitive interface for the workers and managers to manage operations on the farm. The system gives only two options to the end-user that is: The Manager who oversees all farm activities and the Employee who is assigned a task to work on the farm. The farmer in the system has many tools to enable him to manage the farm with keen attention while the employee is just assigned a task and sees to it that it is done. Now in the Manager dashboard, the manager can have access to the Live Feeds, Disease Detection, Workers and Task tools. The farmer could use this functionality by just adding an image to the system and then it detects the possibility of the diseases and also detects how healthy the cocoa pods are. However, the few images we used to train our model gave precise predictions and recommendations. The images captured and used for system testing were good for the drone camera due to the picture resolution of the mobile we used.
Project Title: Uterine Contraction Detection System using Artificial Intelligence Technique
Group Member: Robert Kwame Yeboah & Nketsia Isaac Cromwell.
Supervisor: Dr. Godfrey A. Mills
Abstract:
During the later stages of pregnancy, women go through labour to be able to give birth. One way of observing labor progress is through the monitoring of uterine activities. Uterine contractions become more frequent, last longer and more intense as labor progresses. Health personals find it difficult to monitor and track the progress of labour.
This project seeks to provide an automated system that will monitor and classify the uterine contractions and notify the health personnel when there are some irregularities with limited human intervention. The proposed system developed provides a way to monitor uterine contraction during labour. The system records and computes the average duration, frequency and amplitude of the uterine contractions. The recorded data is sent to the web server through a wireless communication. An artificial intelligent model as added to help classify records into regular and irregular contractions. A web-based application developed also provides a platform that allows users to input patient details, view classifier results and monitor the progress of labour. The developed system was tested and was able to determine the different parameters that are necessary to monitor the progress of labour.
The application is able to classify recorded uterine contraction results (average duration, frequency and amplitude) and also determine whether contractions are regular or irregular.This project will also be useful for doctors, nurses and other health personals in managing their patients. The system developed will provide enormous benefits to users in the monitoring of uterine contraction and the associated risk and complications during labour.
Project Title: Rash Driving Detection System Using Computer Vision
Group Member: Sedem Quame Amekpewu & Samuel Kobina Obeng Andam.
Supervisor: Dr. Nii Longdon Sowah
Abstract:
The primary aim of this project is to develop an embedded system that employs image and video processing algorithms to detect vehicular speeds and flag vehicles that are moving at speeds greater than the allowed speed limits within specified geographical locations. Vehicles that are flagged for speeding will have relevant details extracted from their number plates and sent to the appropriate authorities (the Ghana Police Service). The final product of this project will go a long way to reduce the number of road accidents that plague our roads by serving as a driver-conduct regulatory mechanism on our roads.
According to City News Room, the government of Ghana loses an approximate amount of $233M every year due to road accidents. It is clear that these road accidents do not only cost human life but also substantial amounts of money.
This project is aimed at building an embedded system that will be placed on our roads to take frequent video of cars that pass within the line of sight of the camera. These videos will initially be pre-processed and passed through a computer vision algorithm that detect the speed of each car. If a vehicle is found to be speeding, the algorithm extracts the number plate of the car in question and sends it to an API (Application Programming Interface) managed by the appropriate government body (the Ghana Police Service). The AP then applies monetary sanctions to the registered owner of the car which he/she has to pay in order to be able to renew important document such as roadworthy license and driver’s license. The system will be hereafter referred to as “Maakye wo”, a Fante phrase that means “you’ve been caught”.
The coronavirus pandemic rendered it possible to obtain some needed part for the system and as such the hardware system could not be fully realized. Upon testing of the system, we were able to flag to speeding vehicles and charge the registered owners. Further work can be done on this in the future. This could be the inclusion of more traffic rules such as lane jumping, wrongful overtaking and disregard for traffic signals.
Project Title: Home Management System Using IoT and Artificial Intelligence
Group Member: David Kyei & Faustina Boatemaa
Supervisor: Dr. Godfrey A. Mills
Abstract:
Increasing energy usage and shortage due to the high demand of electricity and devices has presented new challenges for a Home Energy Management System (HEMS). HEMS is a demand side solution efficiently controls energy demand and utilization in the home.
Our project seeks to provide home owners with a system that gives them feedback on energy consumption in individual circuits of the distribution board and individual appliances. It also provides them with a way of remotely controlling energy usage in the home. Our project seeks to provide home owners with a way of knowing how to utilize energy efficiently to reduce cost of electricity usage.
This project delivers a home energy management system (demand side solution) that utilizes artificial intelligence in the form of genetic algorithms and Internet of Things to efficiently manage energy consumption. The proposed system will be developed using a microcontroller based power monitor connected to the distribution board with smart sockets being interfaced to the sockets in the various rooms and an Android application that will allow users to monitor and control their home appliances. The microcontroller interfaced with a distribution board in the home is equipped with voltage and current sensor circuits to monitor voltage and current on the various subsidiary circuits. This unit will ultimately calculate the power usage in the various subsidiary circuits. It will have relays for controlling the various subsidiary circuits. Power consumption in the various subsidiary circuits will be determined and sent via Wi-Fi from the microcontroller through the ESP8266 server. A smart socket unit designed with a current sensor and a Wi-Fi module also measures the power consumption by any appliance connected to it and transmits the power consumption data to our server. The server will hold our database where all power consumption data are stored. This stored data can be viewed graphically with the help of an android mobile application that will be developed. The energy consumption of the various subsidiary circuits and individual devices connected to the smart sockets will be calculated and displayed by the android application.
At the end of our project we were able to design a system that allows homeowners to perform ON and OFF operation, monitor live power readings, intelligently schedule appliances and ultimately reduce the cost of electricity to over 50% reduction.
The project will benefit power generators by reducing the demand for the resources used in generating electricity. Power consumers will also spend less money on power consumption. AI algorithms designed and coded into the android application, will help the system suggest the best schedules to use at various times, which will be preempted upon the user’s consent. This then saves cost to the end users, producers and power distributors.
The project shows that a mobile device can be used to provide demand side management solution in homes by monitoring, controlling and regulating power usage efficiently, which will reduce the cost of electricity usage. At the end of the project, users of the system were able to reduce the cost of electricity usage.