Project 1: P2P Resource Discovery Protocol for IoT EnvironmentsDesign a P2P discovery protocol that IoT devices can use to find nearby IoT resources (e.g., other IoT devices), connected over WiFi. The protocol should consider limited power budget and constrained computational capability of IoT environments. The discovery functionality should be designed using the successful REST approach (i.e. using RESTful APIs) for interoperability and heterogeneity purposes. Students can inspire their design by the CoAP (http://coap.technology/) protocol -- or even repurpose its open source. Evaluation should be carried out to evaluate the overall design as well as to measure the power consumption profile. REFERENCES |
Project 2: Towards Dynamic Access Control in IoT ScenariosDevelop a dynamic access control mechanism to authorize people's access to surrounding devices (e.g., AC control, printer, IoT lock). The use privilege changes according to dynamic context. For example, user U can actuate AC temperature on office A but can only view temp in research lab B. However, if the professor of lab B coexist, then user U can change the temp. Another generic example is: student S can have full access to a course material, but video materials are made available only for students who were present in class. In all cases, the user/or student does not have to reveal his/her identify for authentication purposes, it should be enough for the user to proof that he/she possesses required credentials (example, a member of the research lab in the first case and registered student in the 2nd case) -- using a zero-knowledge protocol. REFERENCES |
Project 3: Ubiquitous Health Monitoring Using Mobile DevicesWrite an android mobile app that collects vital signs of a patient (through a medical device or rest band) and sends it to a backend server that maintains a database for EHR (Electronic Health Records). The backend database could be implemented using MySQL, SQL server, or preferably influx DB (the best and only specialized engine for time series data). A remote health monitoring facility can access these data at the backend, request a fresh copy of current vital signs directly from the user's mobile device, or change the sampling rate on which data are collected. The patient's mobile sends alerts to the healthcare facility if the patient's vital signs exceed a specific predefined threshold. A physician or an authorized care facility practitioner can set/change this threshold remotely on the patient's mobile device. The physician version of the mobile app allows for data visualization on a certain vital sign over a selected period of time. [Hint] design the app functionality using the Django web framework (https://www.djangoproject.com/) for design modularity and RESTful APIs support. All functionality then can be accessed through a single URI/each and can be accessed from any platform tnough a standard web browser. REFERENCES
|
Project 4: Towards Dynamic Context Management for Smart SystemsContext is a key enabler to smart interactions in IoT environments, where context collection, modeling, management and dissemination lie at the core of any smart application. IoT devices (especially smartphones) capitalize on their sensing capabilities to a variety of real-time context information to support context-aware systems making better decisions. Such context includes device profile (OS platform, device capabilities, embedded sensors, etc.), user profile (e.g., role, activities), environmental context (e.g., location, available networks, etc.). While domain-independent context models support a wider scope, and provide common concepts across different domains, they are too general and inadequate for effective semantic relationship inferences. Domain-specific ontology-based context modeling provides better performance due to the ability to optimize inferences, but it incurs high maintenance and does not scale well in large and highly diverse deployments. This project aims to design and implement an ontology-based context representation model that strikes a balance between expressiveness and complexity. The context model will primarily support domain-independent context representation to enable interoperability and serve the general interest of IoT scenarios, but will provide extensibility for including domain-specific ontology and context hierarchy. This extensibility feature aims to support logic reasoning over high-level semantic relationships. REFERENCES
|
Project 5: Building the Resource Management Infrastructure for Community-based IoTThere is a growing interest in community-based IoT deployments, in which a variety of devices (e.g., smart vehicles, smartphones, dash cameras, traffic cameras, etc.) participate to fulfil on-demand sensing requests (e.g., traffic updates, air quality in a certain geographical region, etc.). When a sensing request is received by the system, it matches all applicable sensors, based on their capabilities, to resolve the request. A key attribute to the success of such a deployment is to have the required critical mass of participating devices ready to contribute. To enable this paradigm, we need to maintain an active list of all devices along with their sensing capability to be able to perform runtime matching. This project will develop a platform that provides this physical resource infrastructure to support resource allocation mechanisms. The platform will expose the registration functionality as a RESTful API (using Django) so that mobile users of any platform can use the same API and register their resources. The platform will be able to communicate with the registering device to read all applicable sensing capabilities of the device. The registration process will extract these capabilities and display the results to the mobile user to confirm. The user can just hit “accept” or choose to unregister some capabilities based on preferences. By the end of the registration process, the system will store the device profile including all sensing capabilities and possible user preferences over each one (such that I enable my dash cam to participate if I am in the downtown area and out of rush periods). REFERENCES |
Project 7: Parking Made Smarter with Real-time UpdatesParking is a major challenge for people in urban settings, specially around city centers and business districts. Recent studies show that 30% of the traffic in congested areas is caused by cruising drivers (i.e. drivers looking for vacant parking spots) [ Nawaz et al., 2013]. One solution to this problem is to instrument cities with parking spot sensors, but this comes at a high cost. Another solution is to leverage mobile payments systems to detect when a parking spot becomes empty. However, many users leave before meter expires, leaving a spot falsely occupied. This project aims to design a system that can leverage mobile payments to effectively manage parking spots and detects when a spot become vacant in real-time. You need to design an algorithm to detect when a user return to his previously-paid parking spot and drives away. The project by default will maintain and monitor parking spots and provide accurate real-time updates to drivers on empty spots based on their location. Drivers will be able to use their mobile app interface (or web-based) to hold an empty spot of interest for a limited time before it is paid. The system will be able to suggest a list of relevant vacant spots to the user based on the destination. REFERENCES
|
School of Computing and Informatics, University of Louisiana at Lafayette
All contents copyright © 2018, Khalid Elgazzar.
All rights reserved.