Science, Technology, Engineering and Mathematics (STEM)

Permanent URI for this community

http://hdl.handle.net/20.500.11988/2

Browse

Browsing Science, Technology, Engineering and Mathematics (STEM) by Title

  • Item
    An affordable density-based traffic management system
    ( 2019) Odero, Margaret Anyango
    Traffic management systems at road intersections of most African cities is based on fixed timings where equal length of green light is assigned to each lane at the intersection. This method of traffic management is inefficient because it causes unnecessarily long waiting times for vehicles at the intersections especially during periods when some lanes have few or no vehicles passing through them. In this project, an affordable density based traffic management system is designed and implemented. This is achieved by selecting components that are low cost, consume low amounts of power and do not require digging onto the road pavements. A pi camera with night vision capability is used as the vehicle detection device for every lane. Image processing is then done using an algorithm running on a raspberry pi to determine the number of vehicles on each lane. This information on the number of vehicles is sent to a central raspberry pi via a radio transceiver which then uses an algorithm to determine the next lane to receive green light, as well as the green light duration. In testing the system, the best position of the camera relative to the lane is investigated in an experiment. The furthest distance the camera can be mounted is determined to be 30m. The night vision capability of the camera is also tested. Despite the advantages that come with this low-cost system, it has some limitations. The camera quality is low hence the image processing algorithm may not be completely accurate in distinguishing between objects in the image. The camera’s field of view is also small, necessitating the use of an extra lens that will increase its field of view.
  • Item
    Affordable desktop computing using low end hardware like Raspberry Pi and Cubiebaord NII LOUS NELSON
    ( 2014-04) Nelson, Nii Louis
    The field of computer science over the past two decades has witnessed a huge leap in the areas of improved systems design, hardware processing power and the development of software that maximizes these advances in systems. This leap can attributed to continuous innovation that drives the field to new heights at an ever speedy pace. The pace at which the field is growing requires educators to be abreast with new and changing technology. It is the desire of educators to have access to physical laboratories and computing centers in which they can physically train students. The challenge faced by educators in executing this task is the high expense that comes with the construction of physical laboratories. This write up stems from a project that aims to tackle this challenge by providing low powered, low cost, portable computing systems for students, laboratories and computing centers using devices such as the Raspberry pi and Cubieboard. Although not a perfect system, it promises to be a desirable solution that can be adopted across a wide range of institutions.
  • Item
    Affordable digital guitar effects pedal for local guitarists.
    (Ashesi University, 2019-04) Hazel, Victor
    Music has been a mode of communication and a source of identity for many cultures, especially in Africa. The sound from the guitar is one of the many components that make up music, The electric guitar is a musical instrument, usually made up of six strings which is connected to amplifiers to produce sound. Technological advancement has led to the creation of Guitar effects pedals. However, manufacturers keep making new models of guitar effects pedals with a continuous increasse in cost. It is therefore difficult for low-income young guitarists to have access to guitar effects pedals. This project seeks to use a general purpose microcontroller to create a very affordable guitar effects pedal.
  • Item
    Affordable solutions for computers in schools: Operating system and tools on a USB flash drive
    ( 2014-04) Agoe-Ashiakwei, Ebenezer
    Over the past few years, there have been numerous developments of more efficient and portable ways of using computers. Hardware and storage have been designed to provide progressively smaller systems for computing while speed and data performance has greatly increased. The purpose of this project is to provide a complete Operating System and tools on a USB flash drive. This is done with the aim of providing users with a much more portable, durable and affordable outlook to computing. Users can also share or reuse the same hardware while keeping their operating systems and data separate. This project is necessary as other options require people to have and carry their own hardware, or rely strictly on the operating systems preinstalled on existing machines. Customers have preference for portable and cost-effective devices, while administrators want to simplify asset management. Overall, this project offers a new outlook on using IT hardware while bringing choice and convenience to end users.
  • Item
    An alcohol detection system for vehicles
    ( 2020-05) Yelipoie, Edna
    Drunk driving is a contributing factor to road accidents in several parts of the world. Road accidents caused by drunk driving are usually fatal and have an adverse on the victims and the government, in the long run. Government policies have been implemented to deter individuals from indulging in drunk driving. Positioning police officers with alcohol breath analysers at vantage points on the road has also employed to curb the menace of drunk driving. However, these strategies have barely been able to reduce the rate at which road accidents caused by drunk driving occur. In this project, alcohol breath analysers have been integrated with the ignition system of vehicles to prevent drunk drivers from moving their vehicles.
  • Item
    An algorithm for multi tempo music lyric transcription
    (Ashesi University, 2018-04) Amoah, Hector
    This paper documents an attempt to create an algorithm for multi-tempo music lyric transcription. This paper reviews music information retrieval as a field of study and identifies music lyric transcription as a subset of the music information retrieval field. The difficulties of music lyric transcription are highlighted and a gap in knowledge in the field is identified. There are no algorithms for music transcription that are applicable to all forms of music; they are usually specialised by instrument or by genre. The author attempts to fill this gap by creating a method for multi-tempo music lyric transcription. The methodology used to achieve this goal is a three-step process of taking audio as input, processing it using the REPET separation technique, and transcribing the separated audio file. The result of this paper was a relative success, with the music being separated successfully and the lyrics being transcribed but with accuracy lost.
  • Item
    Analysis of data mining techniques and algorithms for healthcare application using cervical cancer as a case study.
    (Ashesi University, 2018-04) Gouanfo, Cynthia Naela Priscilia Ngaffo
    Cervical cancer is the most common cause of cancer among African women. It is a preventable disease and can be treated if identified at early stages. Given the lack of adequate health care services and the costly nature of colposcopies in Africa, it is difficult to get an early diagnosis. The development of smartphone-based diagnostic tools like MobileODT – with which pictures of the cervix are taken and sent to doctors for diagnosis – promises to address the expensive nature of colposcopy and Pap test; still, the diagnosis of these images is prone to human errors. This project aimed to recommend an algorithm that best classifies cervical images into cancerous and non-cancerous, in order to aid medical officials to give a better diagnosis. K-Nearest Neighbour (KNN), Convolutional Neural Network (CNN) and Support Vector Machine (SVM) were analyzed and compared based on their classification accuracy, sensitivity and specificity and how these results varied after applying Principal Component Analysis (PCA) on the dataset. KNN, CNN, and SVM models obtained classification accuracies of 68.75%, 83.3%, and 66.37% respectively while PCA-KNN and PCA-SVM models had classification accuracies of 78.12% and 62.7% respectively.
  • Item
    Applied project report on the development of the online examination system
    ( 201004) Asiedu, Jonathan Amos
    Examination writing in Ghana seems to be a location based process. This is so because most examinations taken, be it school exams or other external examinations, are taken at specific selected locations. If exams are to be taken in a school like Ashesi University for instance, there is a time table drawn which includes a course, date, time and place of the exam. The processes that precede it include typing and photocopying of questions which can use up a lot of paper, time and space even for a relatively small school such as this. For larger societies or institutions where space is an issue, examination writing can get very uncomfortable since people are crammed into one examination room. This can even lead to cheating on a large scale with after effects such as examination cancellation, grade cancellation, and etcetera. A lot of this can be reduced by developing computer based online examination software systems for use by the institutions that need them.
  • Item
    Ashesi network traffic analysis
    ( 2017-04) Shahadu, Husein
    The web has evolved from an information exchange system to a data mining, knowledge creation or Knowledge dissemination platform where the internet is viewed as a critical and vital component of success by students, lecturers and researchers in higher institutions particularly within universities and colleges. There is thus pressure on network directors managing network services within these institutions to provide regular and correct utilization of the available bandwidth. Network directors are being pressured to ensure that there is sufficient bandwidth available for every network user and also ensure the bandwidth is used productively. With the expansion of digitally made contents and Internet computing demands in the last couple of years, network users often complain of insufficient bandwidth available to completely satisfy their wants. In an attempt to attend to users complains, network directors ought to determine the main objective or purpose of the offered bandwidth and identify unproductive applications eating the bandwidth. In this paper, an attempt is made to investigate network traffic on the Ashesi University network system and to prioritize applications on the network based on user needs and based on what the bandwidth is purchased for. A policy framework is additionally outlined for best utilization and management of the university`s network system. The findings of the study indicate that there is the need to prioritize applications on the network because it was made clear that users access certain types of applications during school hours and some other type of applications after school hours. There is also the need to install certain servers locally in alternative to scaling back the traffic caused during peak days and also during peak hours of the Ashesi network system.
  • Item
    Ashesi Premier League application
    ( 2013-03) Osei, Huberta
    The growth of technology has seen most institution developing ways to reach a wider range of people. The soccer industry is not left out in this race. Most leagues and clubs have come up with different ways to reach their supporters and enhance engagement. One of such ways is the development of applications. The development of these applications does not only enhance user engagement but provides a convenient way to reach soccer fans. The Ashesi Premier League (APL), an upcoming league for the ashesi community can also benefit from this technological growth in the soccer industry. This application will help disseminate information about the league to the community. It will also promote student engagement through predictions and allow students express their views on various issues. This is therefore necessary for the application to be easy to use by the students and serve the different classes of students in the community.
  • Item
    AshesiBot: An automated chatbot for handling Ashesi Facebook Messenger platform
    ( 2019-04) Katena, Primerose Ruvimbo
    As each day passes, new ways of improving our lives through technology emerge. Companies keep on exploring ways of reaching out to their customers and other stakeholders in a way that is more comfortable and efficient. Regardless of their exploration, the problem of presenting information in a way that allows quick retrieval of information by users remains a challenge. To solve the problem of information retrieval, companies have implemented search bars on their website and other functionalities. Unfortunately, search bars have proven to be inefficient because they bring everything that matches with the input of the user. Additionally, others have implemented live chat systems to help mitigate the problem of information retrieval. However, live chats are time-consuming especially when there are few workers behind the live chats systems providing services to numerous customers who demand immediate attention. These problems necessitate a novel approach for retrieving information that meets the needs of businesses and individuals; using chatbots. The idea of chatbots is not new, but it is becoming more refined and useful with the application of machine learning to solve the problem of information retrieval. The advent of chatbots on websites and social media platforms is fast spreading, helping with a 24/7 retrieval of information. The use of automated chatbots to handle a conversation with a business’s customers is now a trend because it allows for quick retrieval of information with little or no support from the staff members. Use of chatbots gives room for staff members of an organization to focus on more significant issues. Customers equally benefit from using automated chatbot system as they no longer need to be in a queue. To sum up, this paper seeks to solve the problem of information retrieval for Ashesi Facebook messenger subscribers using an automated chatbot.
  • Item
    Assessing the impact of (Nigerian, Ghanaian, Zimbabwean, Rwandan and Kenyan) accents on English-speaking speech recognition systems
    ( 2019) Tawo, Ekab-Osowo Samuel
    The use of electronic devices such as computers and mobile phones have rapidly increased over time. Technological powerhouses have inculcated some form of speech recognition in their devices and as such, the increase in usage of this devices provides the increase in interaction between humans and speech recognition. Speech recognition involves adequately transcribing spoken sentences or words into text. This interaction is present all over the world, Africa inclusively, and the continent with its diverse languages comes with its diverse accents and pronunciation of words. A successful interaction between humans and speech recognition involves these systems adequately transcribing the spoken word to the correct text. However, with these transcriptions, errors are prone to occur due to the fact that, mismatch in accents can increase the error rate of a speech recognition system by more than 100% [16]. This Thesis explores to a very large extent the impact that certain African accents such as Nigerian, Ghanaian, Zimbabwean, Rwandan and Kenyan have on the accuracy and transcriptions made by English-speaking speech recognition systems. The findings made in this study states that to a very large extent, the various African accents used in the study does affect the accuracy of the speech recognition systems used for the study.
  • Item
    Attendance management system
    ( 2017-04) Bota, Adwoba Nenge
    Attendance is an important element in face to face education. A lot of schools in advanced countries have some form of attendance management system in place to track student attendance, monitor deviances and a source for recording data. There are many methods used to take class attendance, the traditional method being the use of paper or calling out of student names. Some universities, however, use the student ID card. Over the years, people have built applications or devices to take attendance in class. These methods provide better security than the traditional methods. Some methods are better than others and are worth pursuing. In our local context, universities still depend on papers or roll calling to take student attendance. This project seeks to provide a faster and simpler way to take attendance in class with the use of a QR code. The system should combine speed with accuracy and effectiveness. Keywords
  • Item
    Augmented reality Capture the flag
    ( 2014-05) Hussein, Wumpini Alhassan
    The advent of mobile gaming brought about a new age of how video games are regarded in the context of mobility. Previously, a person would have to own a specialized gaming device that enabled them play on the move. Devices such as the PlayStation Portable and Nintendo’s series of GameBoy devices made this a possibility. However, mobile games have always offered a constraint on the type of interactions they offer players. A mobile game is only as enjoyable as the quality of information it displays on the screen and how it displays this information. A new way of viewing the game world would create a new type of interaction for players, and create a new way for them to enjoy mobile games. This paper reports on the project Augmented Reality: Capture the Flag, which offers a solution to the problem of mundane interactions by using augmented reality as a means of extending the screen of a mobile device. Augmented reality will attempt to reconcile the real world with the game world, creating an illusion of the game world existing in the physical world.
  • Item
    Autograding and detecting plagiarism in student programming assignments
    ( 2019-04) Waweru, Alex
    In computer science, practical assignments ensure that students put the theory they learn in class into practice by writing computer programs to solve problems. Practical assignments also play a critical role in assessing students’ understanding of course materials. For course facilitators, grading programming assignments is a time-consuming task. The course facilitators must run each student’s submission. Moreover, some students copy the code from their friends and change the lexicon and structure. This makes it nearly impossible for the course facilitators to detect plagiarism. A possible solution to these problems is a system that allows course facilitators to write tests that apply automatically to all students’ submissions and consequently allocate grades based on test results. To curb the plagiarism issue, the system should have a component that calculates the peer plagiarism index and flags students’ submissions that may have plagiarism issues. This applied project is an attempt to develop, test and evaluate such a system. While designing the system, it became apparent that running students’ submission and instructors’ tests on the server would pose a security threat to the server. After evaluating possible workaround for the issue, we decided to run the submissions and tests on a docker sandbox within a virtual machine. The plagiarism index is calculated by quantifying the lexical and structural similarities. To integrate the two components, we developed an API. To test and demonstrate the workings of the system, we developed a frontend client to consume the critical endpoints of the API. This project is proof of concept that the solution for the problem can be developed and successfully deployed.
  • Item
    An automated baby monitor for deaf parents
    ( 2020-05) Koroma, Sanuratu
    An automated baby monitor for financially challenged deaf parents in Africa. This system is operating on a 433MHz RF module to aid wireless communication. Its input is from a pulse/heartbeat sensor, which checks the beat per minute (BPM) of the baby when it is higher than 145BPM, indicating that the baby is crying. The Microcontroller used is ATMega328P. The focus of this system is to help a deaf parent recognize that their baby is crying when the parent is not in plain sight of the baby or the parent is sleeping. Hence, this system has two devices, one of which will be worn by the baby and another which will be worn by the parent, and communication is wireless. This system is made from affordable electronics components to ensure that the final product is cheaper as compared to its competitions.
  • Item
    Automated collection and visualization of road quality data to aid driver navigation
    ( 2017-04) Boohene, Kwabena Gyang
    In nations worldwide, roads play a major part in the livelihood of individuals. They aid in the transport of both people and goods from one location to another. In developed countries, most roads are tarred making travel easier and quicker. In developing countries, this is not the case. Quite an amount of road work and development is still needed in these regions. These roads normally have different grades of road quality and have different effects on the individuals that use them. With the introduction of mapping applications such as Google Maps, individuals can place information about various locations on a mapping interface. In this project, an application is built to allow individuals collect data about the quality of roads around them. This data will be viewed on a map interface. With such data, drivers would be able to choose they type of roads they would want use.
  • Item
    Automated course scheduler for Ashesi University College
    ( 2014-04) Agbenewosi, Jennifer Fenteng Valentina
    Ashesi University College is faced with the challenge of effectively scheduling courses at the beginning of the semester so that there are no class clashes for both lecturers and students. In an attempt to solve the Course Timetabling Problem at Ashesi University College, five algorithms: Genetic Algorithm, Constraint Programing, Particle Swarm Optimization, Simulated Annealing and Tabu Search algorithm, which are known for their use in solving University Course Timetabling problems have been studied and based on their ease of implementation, their robustness in arriving at feasible solutions, their computational speed and whether an optimal solution is always guaranteed, Particle Swarm Optimization algorithm is chosen to implement a solution to the Ashesi University Course Timetabling problem. This project is focused on eliminating course conflicts and creating an optimal table based on teachers‟ preferences for certain timeslots to teach during the week. The paper outlines the assumptions and steps including explanations on Particle Swarm Optimization used in constructing the timetable base on teachers‟ preferences. Test conducted on the project proved that the use of Particle Swarm Optimization to solve the Ashesi Course Timetabling problems is in the right direction.Finally, the paper proposes a focus on other areas of the course timetabling problem at Ashesi University College, using the same Particle swarm optimization procedures described in the paper to help provide a complete solution to the timetabling problem of the school.
  • Item
    An automated email classification system for the Ashesi Support Center
    ( 2019-04) Boahene, Kwame Owusu
    The widespread usage of the internet has made email an indispensable tool for communication within organizations. Today, email is used by support centers as one of the mediums for providing solutions to the daily internal problems’ organizations face. An example is the Ashesi Support Center which is the hub for solutions for all problems and questions relating to IT, facilities, logistics, and other issues on the Ashesi University campus. In dealing with problems, the Ashesi support center classifies emails as either an IT related issue or an operations related issue. However, the support center does not have a way to automatically classify the emails. Hence, a support personnel manually sifts through the emails to group them. This can be a cumbersome process considering the support center receives over 40 emails daily during peak periods. Harnessing the power of machine learning, a classification model is built to automatically group emails the Ashesi support center receives.
  • Item
    An automated plastic bottle collection system
    ( 2020-05) Amarteifio, Wendy Amarteley
    Improper disposal of refuse has in recent times caused harm to live and properties. In our cities and towns, it is common to find waste at inappropriate places such as gutters. This project focuses on the use of Internet of Things and smart systems to aid the waste management system, by providing an automated plastic bottle collection system. This project is focused on providing an efficient method of automating plastic bottle collection by placing value on all bottles. Value is placed on plastic bottles returned to motivate individuals to recycle; vouchers are generated per each bottle returned. The prototype of the automated plastic bottle collection monitors number of bottles, bin level and generate vouchers. Information on number of bottles and bin level are secured in a web application to determine when bins should be emptied