This chapter presents an integrated framework for transforming urban transportation into a smart, adaptive ecosystem by combining three key technologies: vehicle-to-vehicle (V2V) communication, large-scale mobile crowdsensing, and autonomous software agents. Urban centers today face chronic congestion, safety risks, and environmental pressures that traditional traffic controls—fixed signals and centralized dispatch—cannot effectively address. Vehicular communication creates a real-time mesh in which vehicles exchange safety and status messages at millisecond latencies, enabling collision warnings, cooperative platooning, and dynamic routing. Crowdsensing leverages the ubiquity of smartphones, in-vehicle sensors, and IoT devices to gather city-wide data on traffic flow, road conditions, air quality, and more, yielding a macro-scale picture of urban dynamics. Autonomous agents, embedded in vehicles and edge servers, act as intelligent intermediaries: they filter and fuse raw sensor streams, make split-second safety decisions, negotiate cooperative maneuvers, and enforce privacy and trust policies. We begin by defining the fundamentals of V2V networking and crowdsensing paradigms. We then describe the agent architecture—perception, decision, communication, and learning modules—and how agents operate across device, edge, and cloud tiers to achieve low-latency responses and scalable coordination. A three-layer system architecture is detailed: the device tier handles local sensing and immediate control; the edge tier aggregates data from multiple vehicles for real-time analytics and localized actuation; and the cloud tier performs deep learning, long-term planning, and digital twin simulations. Advanced data processing techniques are essential to convert noisy, heterogeneous streams into actionable intelligence. We explore sensor fusion methods (Kalman and Bayesian filters), machine learning models for traffic prediction and anomaly detection, real-time stream-processing pipelines, and privacy-preserving analytics such as federated learning and differential privacy. These methods enable both micro-level safety interventions (emergency braking, blind spot alerts) and macro-level optimizations (signal timing adjustments, city-wide routing strategies). Through case studies—from smart intersections in Los Angeles to air quality mapping on public transit—we demonstrate how the integrated framework reduces travel times by up to 20%, cuts intersection delays by 25%, boosts hazard detection accuracy, and delivers high-resolution environmental monitoring. Applications in dynamic parking guidance, emergency response corridors, and transit fleet optimization further illustrate the platform’s versatility. We identify critical challenges—data quality and trust, privacy and ethics, interoperability across V2V standards, resource constraints at the edge, cybersecurity vulnerabilities, and public engagement—and outline research opportunities in trust management, edge-cloud orchestration, explainable agent behavior, and multimodal mobility integration. Emerging trends such as 6G-enabled V2V, large-scale digital twins, and AI-native networking promise to extend these capabilities. In conclusion, by harmonizing V2V communication, crowdsensing, and autonomous agents, urban transportation systems can evolve from static, centrally controlled infrastructures into dynamic, self-optimizing networks. This paradigm shift offers the prospect of safer roads, smoother traffic, lower emissions, and more responsive public services—paving the way toward truly intelligent, sustainable cities of the future.