Design & Development of Smart Electric Vehicle Safety Device by using IoT and AI

As the usage of electric vehicles grows in popularity, the risk of accidents caused by them is also drastically increased. High speed, drink-driving, distracting thoughts, overstress, and malfunctioning electronic gadgets, over the temperature in charging are the leading causes of accidents & fire. Also, in electric battery-operated cars, if the battery charging mechanism is not properly thermally stabilized, the battery may catch fire. As a result, the proposed system deals with accident detection systems that occur as a result of a driver's carelessness or system failure while operating an electric vehicle. Proposed Device created with a cardiac sensor in the seat belt and vehicle along with battery temperature, vibration & accelerometer sensors which can help to analyze the vehicle's yaw, roll, pitch, and slip motions. All of the data collected is a synced to ESP-32 controller through IoT-based cloud storage analysis to notify a neighboring helpline center about the accident. This data is also utilized in an AI-based model to detect one of the sovereign sources of death in growing countries i.e., cardiovascular disease, and it's cited by several throughout the globe. Cardiovascular disease, on the opposite hand, is called heart disease. This consolidates some dangerous factors of heart disease and the desire of the era of promoting correct, reliable and smart approaches that enable early identification to achieve quick treatment for the disease. This analysis paper also presents numerous attributes associated with a heart condition, and also the model is based on supervised algorithms of machine learning such as Support Vector Machines, Logistic Regression, Naïve Bayes, Decision tree, and random forest and algorithms of deep learning such as Artificial Neural Network, TabNet etc. It uses the prevailing dataset from V.A. Medical Center, Long Beach, and Cleveland Clinic Foundation: Robert Detrano, M.D., Ph.D. It includes 303 occurrences and 76 attributes. Only 13 attributes are considered for testing out of 76 attributes. This is important to prove the performance of various algorithms. This analysis paper targets to visualize the likelihood of blooming heart conditions within the drivers with the comparison of traditional algorithms with deep learning techniques and will alert drivers by sending information via personal email, text message, or on the app.