This thesis tries to find a solution for the problem of managing and monitoring the banana shelf in a supermarket using IoT. The research focuses on using a wireless sensor that detects some features of the banana shelf while being non-intrusive. The three main features that are examined of the shelf are the quality and quantity of the bananas and the quality of the shelf. First a research was conducted to find the best sensor to use for these measurements. The chosen sensor is a color image sensor, the platform for the IoT device is a Raspberry Pi. Using the python programming language in combination with the openCV library image processing was used to detect the features. The image is first smoothed using a Gaussian filter, afterwards the foreground is segmented. The different segmentation methods are researched and adaptive thresholding is used. To determine the quantity of the bananas and quality of the shelf the stickers on the bananas are detected. This detection is implemented using different filtering methods ranging from spectral filtering to color thresholding. With the segmented foreground the quality of the bananas is assessed using a color histogram. This information is then sent to a communication module that is connected to a IoT dashboard for user interpretation. With the proposed design the status of the shelf including the percentage of the shelf filled, the quality of the bananas on the shelf and the neatness of the shelf are available for a supermarket manager to better organize his supermarket. This sensor makes it possible to better organize the banana shelf and act preemptive instead of reactive.

During 9 weeks of investigation multiple solutions for implementing a communication module in a supermarket have been explored. These results led to three options which have been further researched. Bluetooth Low Energy (BLE), 433 MHz and ZigBee have been selected as the best communication protocols for sensor to gateway communication. Chips for these protocols have been bought and for BLE and 433 MHz libraries have been developed. For the gateway and the communication to the server a Raspberry PI and Mobile Internet have been selected and these have been integrated. During the measurements we came to the conclusion that for the RFM69HCW (433 MHz) and the BLE Nano (BLE) the sleep mode did not work as specified. Resulting in high power consumption and thus worse battery life (sleep mode consumption for RFM69HCW 3.37 mA, BLE Nano 4.38 mA and for XBee 43.2 휇A). The final conclusion of this thesis is that, with the current research, XBee (ZigBee) is the best communication protocol for communicating within a supermarket. But in further research BLE Nano and RFM69HCW can perform better if the sleep mode is improved to work according to what is specified.

Internet of things (IoT) applications become more and more prominent in our modern society. IoT has the potential to improve business processes and change the ways we live. KPN New Business has a high affinity with IoT projects and supplied the problem definition that this project builds on. The project has been done in close collaboration with KPN. This document describes the design process for a stand-alone obstruction detection system for supermarket environments. The focus lies on obstructions near emergency exits, where they pose safety threats and obstructions in shopping aisle, where they hinder customers. The system should detect obstructions and communicate the presence of an obstruction to a central server. The communication is however not within the scope of this design report. Five sensing techniques were considered for the detection system, ultrasonic ranging sensor were deemed the most suitable for this application. Ultrasonic sensors lend themselves well for IoT applications as they are easy to implement, cheap and operate on low power. A MaxBotix MB1300 sensor was chosen for the implementation of the prototype as it offers an Analog Envelope (AE) output of the measurements. This allows for multiple objects to be detected in a single measurement. Furthermore, the MB1300 sensor has the largest beam width in the AE product line and can thus cover the most surface area. The processing of the measurements in the prototype is done using an Arduino Mega 2560. The Arduino board allows for easy prototyping and has sufficient memory on-board to store and process the measurements. The measurements are sampled using the built-in ADC of the Arduino at a sampling frequency of approximately 8.93 kHz. The sampling frequency is a trade-off between the spatial resolution of the system and the memory required to store the measurements. An algorithm is developed for the detection of objects from the sampled measurements. The process involves the windowing of the measurements to only look at relevant samples. A background subtraction is performed to avoid the detection of scenic objects, and the objects are detected using a threshold. A data structure was created in order to store the detected objects. An algorithm was developed to check whether detected objects remain static. If an object is repeatedly detected it is marked as an obstruction. If an object goes unseen for a certain period, the object is removed from the memory. Thresholds for when an object is marked as an obstructions, the deletion of objects and measurement intervals are suggested for emergency exits and shopping aisle systems. These should however be verified in further development. The performance of the system was evaluated in testing environments for the emergency exits and shopping aisles. Testing results showed that the system performs well on most design criteria. The area covered by a single sensor was however below expectations. This could be problematic for the implementation of the system in shopping aisles, as lots of sensors would be required to cover entire shopping aisles. Also some false positives occurred in the detection of obstructions. This could however easily be solved with minor changes to the detection algorithm. Based on the testing results the conclusion is drawn that the design goals are partially met. The system is able to detect obstructions, but does not yet operate as a stand-alone unit. This design goal was however considered in every design choice. Power consumption of the prototype remains to be tested, this however is also dependant on the communication system. Recommended is to further look into the power consumption of the system, the implementation of a multi-sensor system to cover more floor area in shopping aisles, and to reconsider CCTV images for the detection of obstructions in shopping aisles.