Although many promising measures have been taken to reduce road traffic accidents, they still are responsible for millions of deaths yearly. New measures such as AI’s application in vehicles could play an increasing role. To study AI’s potential for road safety three applications, namely obstacle, traffic sign and cut-in detection are studied in Tesla’s Model 3. The AI behind these applications are presented to highlight how they could circumvent potential road danger. In particular, the application of convolutional neural networks for image analysis is studied in-depth. The shortcomings of AI are highlighted by a recent Autopilot crash, and simulation as an alternative to real-data collection is discussed. The essay concludes that AI will inevitably improve with developments in computing power and hardware, unsupervised learning and pattern recognition. Nevertheless, for enhanced road safety, humans need to stay alert on the road and appreciate AI as complementary support.
Every year, approximately 1.35 million people die due to road traffic accidents, and 20-50 million sustain non-fatal injuries (WHO, 2020). Around 90% of accidents are related to human error, such as inattention, speeding, and improper lookout. Today, numerous safety measures are in place to reduce the rate and impact of accidents, such as airbag, seatbelt, and speeding regulations. However, one measure that could increasingly play a life-saving role, is artificial intelligence (AI) and deep learning found in autonomous vehicles. There are five levels of automation (SAE, 2018). At level 5, vehicles are fully autonomously. Whereas in levels 3-4 vehicle can travel fully-autonomously but require intervention in exceptional circumstances. However, this AI applications in driver-assisted and partially-autonomous cars, since further autonomy levels are currently not permitted in mass-produced vehicle. Thus, the essay aims to tackle how AI can enhance road safety and potentially safe lives in common vehicles and real-life-scenarios. The essay provides a literature review and showcases concrete examples, namely, obstacle, traffic sign, and cut-in detection. Then challenges of implementation are discussed. Lastly, the essay concludes with future directions for this topic. Prior literature leans toward an agreement that AI could enhance road safety. Bonnefon and colleagues (2016) research finds that AI could reduce human and environmental factors leading to accidents. Research by Stanford University found that autonomous vehicles (AVs) with a neural network structure that utilizes sufficiently, and appropriately trained data can make better future predictions than an average driver. Artificial neural networks are algorithms that enable information processing and build the basis for deep learning. These findings were supported by Jacob stein (2019), who points out that AVs have fewer crashes per millions of miles than non-autonomous vehicles. Many prior works focus on applications of AI in level-1&2 cars using computer vision (CV), which refers to the knowledge gained by computers through digital images and videos. State-of-the-art-developments in CV have presented reliable depth-estimation via convolutional neural networks. CNNs are deep neural networks with multiple hidden layers used for visual image analysis. State-of-the-art developments in CNN, such as Faster R-CNN and Mask R-CNN, have reduced the running time of region-based detection networks
However, there is limited literature that studies these three applications concisely in one paper and their usefulness for road safety, including potential challenges. This paper seeks to do so in the following sections.AI is used for collision avoidance through obstacle detection, which can circumvent potential accidents caused by human error. The examples of AI are applied to a Tesla Model 3, where the data is collected from its eight cameras, and twelve sensors (Tesla, 2020). Unlike most a partially automated vehicles, Tesla does not use LiDAR, which refers to light detection and ranging, and is valuable due to its depth knowledge. Instead, it relies on computer vision. The human eye immediately recognizes a kangaroo on the street. A computer, however, sees a million brightness numbers in a grid of all the pixels. When CNNs initially predict the object, the connection strengths between networks are vague, and the prediction will be random. Therefore, CNN training is relevant, and to reduce the error back propagation is applied, which is an algorithm in supervised learning that can adjust the weights and biases of neural networks.
Simply the conclusion is, Artificial Intelligence and 5G will ultimately improve road safety, helping reduce the total number of accidents on the road. In addition, artificial intelligence will add advanced road safety features in cars, such as the ability to understand road signs, auto break and more.
You must be logged in to post a comment.