From Selfish Driving to Social Optimality: Insights from a “Nature” Interview with Professor Christos Cassandras
In the modern urban landscape, driving is fundamentally “selfish.”
Each driver chooses a route based on personal convenience, often leading to traffic congestion. Distinguished Professor of Engineering, Head of the Division of Systems Engineering, and CISE Faculty Affiliate Christos Cassandras (ECE, SE) is proposing a solution to this problem—transforming cities into integrated “cyber-physical systems” in which vehicles no longer compete but cooperate.
In a recent Nature report, titled “Smart cities drive into the future,” Cassandras was interviewed and shared his proposal for a networked environment in which every vehicle shares data with another vehicle or a coordinator regarding position, velocity, and destination to optimize routing across the entire city.
At the core of Cassandras’ work is the transition from “user-optimal” behavior, drivers acting in their own best interest, to “system-optimal” outcomes. To bridge this gap, his research focuses on Connected and Automated Vehicles (CAVs). Unlike traditional autonomous cars that rely solely on onboard sensors to “see” their surroundings, CAVs “talk” to one another and to city infrastructure, such as traffic lights.
Cassandras says that if computers manage the driving, the human errors responsible for an estimated 94% of accidents could be drastically reduced. “Computers don’t sleep. Computers don’t blink. They don’t drink,” Cassandras said.
In this model, CAVs use a digital ledger to execute maneuvers. For instance, a car that allows another vehicle to merge in front of it, a move that benefits overall traffic flow but slightly inconveniences the individual, could earn micropayments in a virtual currency. This incentivizes “pro-social” driving behavior, applying economic incentives to clear physical traffic jams.
However, Cassandras acknowledges that the hurdle isn’t just the code, but preventing “smart users” from gaming the system to accumulate cash without contributing to the common good. In the report, he said that users could intentionally create situations with other drivers that would result in micropayments.
Beyond safety and speed, Cassandras advocates for a radical rethinking of urban real estate.
Available statistics and an overall urban traffic analysis suggest that if cities transitioned to shared “robotaxis” that are in constant use, the need for parking would plummet. Since most private cars sit idle for 95% of their lives, freeing up that space could enable cities to convert asphalt parking lots into high-value community assets such as hospitals, schools, and parks.
Ultimately, Cassandras emphasizes that the basics of this technology exist, and research is underway towards the “internet of cars.”
The true challenge, though, is political and economic. The future of the smart city depends on whether governments and manufacturers can agree on shared protocols. For Cassandras, the goal is a “seamless moving web” in which data, not just pavement, dictates vehicle flow.