Vehicular Packet Distribution via Predictive Roadside Caching
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Vehicular network environments face challenges in maintaining quality of service due to high variability and availability of wireless links caused by vehicle mobility, leading to temporary network disconnections and reduced quality of service.
Innovation Solution
A system and method that utilizes distributed content caching, forward error correction, network coding, and packet spraying to multiple access points to ensure reliable data delivery, predicting vehicle paths and pre-caching data at access points along the route to minimize data transfer time and reduce packet loss during handoff events.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Loss of energy
If data is transmitted in real-time from remote servers to moving vehicles, then network bandwidth is utilized efficiently, but packet loss increases due to vehicle mobility and wireless link variability
Solution Approach 1:
The system predicts vehicle trajectories and pre-fetches data to roadside access points before vehicles arrive. This preliminary action ensures data is cached at optimal locations in advance, eliminating the need for real-time transmission and reducing packet loss due to mobility.
Solution Approach 2:
Roadside access points serve as intermediaries between remote servers and moving vehicles. They cache data locally and deliver it to vehicles when needed, acting as a buffer that decouples the reliability requirements from the wireless link variability.
2Reliability
If data is cached at roadside access points to reduce packet loss, then reliability improves, but network infrastructure complexity increases
Solution Approach 1:
The system enables existing roadside access points to perform multiple functions: traditional wireless communication plus intelligent data caching and prediction. This multi-functionality leverages existing infrastructure while adding caching capabilities, avoiding the need for completely new specialized hardware.
Solution Approach 2:
The system uses vehicle-generated location and trajectory data to autonomously predict future positions and pre-fetch required data. This self-service mechanism eliminates the need for complex centralized control systems, as the network adapts automatically based on observed vehicle patterns.
3Reliability
If redundant packets are transmitted to mitigate packet loss, then reliability improves, but network bandwidth consumption increases
Solution Approach 1:
Instead of transmitting redundant packets over the air, the system pre-fetches data in advance and caches it at roadside access points. This eliminates the need for retransmission redundancy, as data is already available locally when vehicles need it.
Solution Approach 2:
The system creates copies of data and stores them at multiple roadside access points along predicted vehicle trajectories. This spatial replication replaces temporal retransmission, allowing reliable delivery without consuming additional bandwidth during active transmission.
4Reliability
If the system predicts vehicle trajectories and pre-fetches data, then data delivery reliability improves, but computational requirements and processing time increase
Solution Approach 1:
The system predicts only the portion of vehicle trajectories relevant to data delivery, rather than computing complete future paths. It focuses computational resources on predicting positions where data will be needed, performing partial predictions that are sufficient for caching purposes without excessive processing overhead.
Data Source
Figure 1
Figure 2
Figure 3
AI summary
A method is provided in one example embodiment that includes receiving a request for data from a source device, such as an on-board unit of vehicle or a mobile device coupled to an on-board unit, over a vehicular network. Location data, such as GPS coordinates, speed, and heading associated with the source device may also be received. A travel path for the source device can be predicted based on the location data, and an access point to the network can be located within range of the travel path. Packets associated with the requested data may be sent to the access point and then forwarded to the source device when the source device is in range.