Improving Object Detection Versatility with 6G in VANETs

After the widespread introduction of 5G cellular communications, the academic and business communities have begun to focus intently on the next generation of wireless communication systems, known as 6G. Vehicular edge computing (VEC) is one of the upcoming technologies that can guarantee the AI algorithms used in 6G networks will work reliably. To address this issue, this work proposes a method of employing 6G VEC to make AI model deployment more reliable, with the object identification job as an illustration. Stabilizing the model and then tweaking it are the two main phases of this plan. The former includes supplementing the model with state-of-the-art techniques to make it more stable. There is a give-and-take between model performance and runtime resources in the latter because of the targeted compression strategies used, which are model parameter trimming and knowledge distillation. The numerical findings show that the proposed method may be easily implemented in the onboard edge terminals, where the presented trade-off outperforms the other known solutions.