Link Lifetime Approximation in Vehicular Communication using a Comparative Movement and Vehicle Control Maneuver-driven Technique

Abstract

High mobility and rapid topological changes are characteristics of vehicular networks. As a result, in contrast to other networks, the connections between the automobiles are erratic and only last a brief period of time. Therefore, understanding the connection lifespan is crucial for creating effective communication between vehicles without packet losses. Only sensor-based measurements have been taken into account in previous efforts in this field to forecast link lifetimes. However, as they lack knowledge about the vehicle's expected future behavior, we anticipate that estimating link lifetimes solely from sensor data may result in less accurateestimates. We suggest using throttle and steering angle driving action outputs in conjunction with brake components to improve sensor readings and provide more futuristic relative motion information in order to address this issue. In particular, we use throttle changes to calculate jerk and integrate and combine the sensor signals to calculate average acceleration and velocity values. We then calculate new motion components, taking into account the steering angle change when it varies from the prior timestep. We suggested modeling the connection lifetime prediction problem through optimization, taking into account the relative velocity of the automobiles,adding jerk, and modifying with driving outputs. But because of that method's high computational cost, we also suggest a suboptimal method based on deep neural networks (DNN) to lower thecomputational difficulty. The proposed model is simulated utilizing NS3 for vehicular data transmission and CARLA for self-driving by leveraging a pre-trained driving model. The findingsdemonstrate that, in comparison to current methods, the link lifetime forecasts of the suggested models are substantially closer to actual link durations; as a consequence, the suggested method may be applied to enhance vehicular communication.