Autonomous vehicles
30 seconds (Stewart 2018). Naturally, this would only be a fraction of the data produced by a self- driving vehicle due to the extra hardware being used. This data would need to be combined (through sensor fusion), sorted, translated into a form that the computer can read, and then instructions need to be made on how to deal with the findings. Sensor fusion is when the computer combines all the data received from the different, independent sensors to create a clear picture/ outcome of the world around the car. In other words, sensor fusion is responsible for object recognition and classification. This function is crucial in the decision-making process. Even if the hardware on the car is revolutionary, and even if the algorithms have a solution for every possible outcome, the wrong decision will be made if a clear and decisive picture of the surrounding environment is not established. However, obtaining this clear representation of the world can become problematic when individual sensors recognize the same object as different things. This is when the ‘best’ outcome is determined and applied. Furthermore, the recognition of objects becomes significantly more unreliable when identifying smaller objects (Thakur 2017).
Infrared Sensors for Autonomous Vehicles. Thakur, R. 2017 The last two rows of this table offer difficult problems: the self-driving vehicle must
determine whether running over a dog presents a higher risk than braking suddenly and causing a collision. Similarly, would the pothole be recognized early enough to avoid, or is it small enough to drive over? One of the most apparent benefits of autonomous vehicles is their ability to ‘communicate’ and connect with one another. Currently, there are 2 means of communication: vehicle-to-vehicle communication (V2V) and vehicle-to-infrastructure (V2I). V2V allows vehicles to communicate with each other and share information, including direction of travel, speed, route, and other characteristics such as the changing of lane. V2I allows vehicles to connect with landmarks and build ings, such as: ‘smart road signs,’ lane markings, and construction zones (Kirkpatrick 2018). However, for this transfer of information to be safe and secure, a wavelength of communication has been established, called ‘Dedicated Short - Range Communication’ or ‘DSRC.’ Naturally, self -driving vehicles need a network to facilitate V2V and V2I, but standard cellular connection is slow, unreliable, insecure, and even negligible in remote areas; this dedicated bandwidth is specifically for automotive use which drastically improves safety and convenience. DRSC can be thought of as wi-fi-it is used for over-air data transmission through V2V and V2I exclusively. However, it does not need a cellular connection, instead, it communicates directly with other vehicles and infrastructure, having a low latency of around 0.02 seconds (Unrau 2020). This means that the delay after an instruction or message has been submitted is roughly 0.02 seconds, while the bandwidth experiences minimal interference. As well as being reliable in varying weather conditions, the bandwidth of DSRC is protected by the Federal Communications Commission (FCC) law.
18
Made with FlippingBook interactive PDF creator