The central aim of this project was to develop an Internet of Things-enabled selfdriving car for a Lego smart city. The goals of this project were to develop a basic non-AI-driven self-driving algorithm based on an online tutorial, an AI-driven selfdriving algorithm controlled by a convolutional neural network of my own design, and a sensor failure response system that can guide the car to a safe emergency stop location if one or more sensors on the car fails. While this car has no direct realworld application and is effectively a toy, this car demonstrates principles and functionalities that a real self-driving car will require when driving. This project operated under a limited scope which only involved development of basic driving algorithms to control the vehicle, training of a convolutional neural network to control the vehicle, and development of a sensor failure response system. The car is controlled by an Arduino Nano 33 BLE Sense microcontroller, which performs sensor fusion using the data provided by a camera, an ultrasonic transducer and two downward-facing colour sensors in order to control the vehicle using two electric motors – differential torque is applied to steer the vehicle. Due to the limited scope of this project, the following assumptions were made: • The car will operate on an empty road with no other traffic • All roads the car will operate on will have a uniform road marking scheme • The sole priority of the car is to drive along the road without veering off the road or colliding with obstacles on the road – navigation to a specified destination is not a concern. The non-AI-driven self-driving algorithm was completed and tested and the CNN was completed, trained and tested. However, I was unable to develop the full AI selfdriving algorithm or the sensor failure response system.