Whilst Thermal Sensors have been around since the 80s, early Automotive thermal sensors used Barium Strontium Titanate (BST), were sensitive to ~100mK, and had a pixel pitch <48um. Today, there are aSi and VOx sensors that allow for improved sensitivity and smaller pitch, with up to 50% reduced size and improved image quality. This presentation will focus on the improvements in Thermal Sensors on the market today, as well as the crucial role of software in nighttime accident reduction. We will outline our project to enhance the classification algorithm, utilizing Convolutional Neural Networks (CNNs) based on Artificial Intelligence (AI). CNNs have proven to be highly effective in identifying shapes and features in camera pictures, thereby enhancing the performance of the system. Through this approach, we have successfully achieved pedestrian AEB capabilities at speeds of up to 90kph, and even conducted rain testing at 60kph without any contact to the pedestrian dummy, aligning with recent NHTSA and anticipated NCAP Mandates.