Contribution boundaries remain explicit
The mobile application and predictive workflow are presented as Nasir’s main contribution; the hardware logger was built by the project partner.
Loading portfolio
Robotics and computer vision / 2022
Intelligent PV Monitoring and Power Prediction
Household PV operators need visibility into solar, battery and load behaviour, while academic forecasting requires a traceable data-cleaning, modelling and evaluation pipeline.
Built the GreenHouse Logger Xamarin.Forms/MVVM Android application and predictive-model workflow. The partner-built IoT logger supplied measurements over Wi-Fi; pandas and scikit-learn prepared solar, battery and load inputs for the report-defined feed-forward network.
- Role
- Android application and predictive-model contributor
- For
- BEng Computer Engineering project
- Stage
- Research prototype
My part
What I built
- Built and emulator-tested the Android application and connected it to the logger over Wi-Fi.
- Cleaned and normalised the supplied dataset and contributed the predictive-model workflow.
- Clearly separated personal mobile/model contribution from the partner-built hardware logger.
Decisions
The choices that shaped it
Academic metrics stay labelled
R² and loss behaviour are report-based academic results, not independently replicated production forecasting.
What has been shown to work
- The report records 1,770 samples, 702 modelling samples, R² 0.998 and loss stabilisation after roughly 30 epochs.
- Evaluation methods included MSE, MAE, MAPE, R² and Pearson correlation.
Current stage and limits
- Academic validation, limited homes and samples, Android-only application and no production deployment.
Screens
From the product
Technical detailOpen the system view
SYSTEM ARCHITECTURE
IoT monitoring and academic prediction pipeline
A partner-built logger captures inverter data, an Android app visualises and exports measurements, and a cleaned feature pipeline feeds the report-documented neural-network evaluation.Solar, battery and load measurements
Partner-built Wi-Fi data logger
Xamarin.Forms MVVM monitoring and export
pandas cleaning and scikit-learn normalisation
256 / 512 ReLU layers and Adam
MSE, MAE, MAPE, R² and correlation
- 01PV systemsmeasurementsIoT logger
- 02IoT loggerWi-Fi dataGreenHouse Logger
- 03GreenHouse LoggerexportDataset pipeline
- 04Dataset pipelinenormalised featuresFeed-forward ANN
- 05Feed-forward ANNpredictionsAcademic evaluation
Security and privacy
- No household-identifying data is published.
- Report metrics remain labelled as academic evidence.
Built with
More work
ARIRA ROS-based autonomous retail robot combining real-time shelf perception, multi-object tracking, autonomous navigation, customer-request handling, and ERP synchronisation.Validated in simulationRobot Guide DogReal-time ROS perception and audio-guidance system developed for a Unitree Go1 guide-dog concept, combining object detection, scene context, and synchronised spoken cues.Research prototypeVision Operations PlatformBrowser-to-inference platform for streaming camera frames to a FastAPI detection service, visualising detections in the browser, and monitoring streaming behaviour in real time.Maintained