Page 822 - AI for Good Innovate for Impact

AI for Good Innovate for Impact

infrastructures. The long-term vision is to seamlessly integrate AI-enabled 6G V2X technologies
into smart cities, enhancing mobility while minimizing environmental impact.

3 Use Case Requirements

• REQ-01: It is mandatory that the system supports wireless infrastructure with ultra-low
latency and high data throughput to enable real-time V2X communication[5].
• REQ-02: It is critical to implement a Deep Reinforcement Learning (DRL) algorithm—
specifically Proximal Policy Optimization (PPO)—to dynamically manage vehicle-to-RSU
assignments.
• REQ-03: It is critical that the DRL agent adapts in real time to network changes such as
traffic density, RSU load, and vehicle mobility.
• REQ-04: It is mandatory to incorporate a retraining mechanism that updates the DRL
model based on evolving traffic patterns and urban layouts.
• REQ-05: It is mandatory that each vehicle be assigned to exactly one RSU at any given
time, ensuring a valid and exclusive communication link.
• REQ-06: It is critical that the system minimizes energy consumption across RSUs while still
satisfying quality of service (QoS) constraints such as latency and bandwidth.
• REQ-07: It is critical for the DRL agent to continuously learn from environmental feedback
to enhance future decision-making and improve resource efficiency.

4 Sequence Diagram

5 References

[1] Li, P., Z. Xiao, H. Gao, X. Wang, and Y. Wang. "Reinforcement Learning Based Edge-End
Collaboration for Multi-Task Scheduling in 6G Enabled Intelligent Autonomous Transport

786