Navigating Mixed Traffic: How Autonomous Vehicles Interact with Human Drivers — and Why It Matters

Autonomous vehicles (AVs) promise safer, cleaner, and more efficient mobility. Yet in reality, the rollout will not replace all human-driven vehicles overnight. For the next several decades, mixed traffic environments — where autonomous cars share roads with human drivers, pedestrians, cyclists, and two-wheelers — will define the future of transport in India, Australia, and beyond.

What We've Learned from Global Data

California's Department of Motor Vehicles (DMV) maintains one of the richest datasets of AV crashes and disengagement reports. A study of this dataset revealed that the conditions influencing AV crash severity diverge significantly from those of human-driven vehicles. For example, time of day, land use type, and the mix of surrounding vehicles are stronger predictors of crash outcomes for AVs than for traditional vehicles.

Other research has gone further. A 2022 Bayesian hierarchical model of crash severity highlighted that mixed land use environments increased AV crash severity, while daytime conditions reduced it — a pattern distinct from conventional traffic.

The lesson: autonomous systems don't simply "inherit" the risks of human drivers — they create new safety dynamics that must be understood before widespread adoption.

Implications for India and Australia

In India, the complexity of mixed traffic is far greater than in the U.S. Buses, trucks, scooters, pedestrians, hand-carts, and livestock often share the same lanes. Without localized modeling, importing AV technology tested in homogenous traffic could create new hazards rather than solve existing ones.

Australia faces a different but equally pressing issue: vast rural highways where AV systems must interact with variable lighting, wildlife crossings, and low-density traffic. In such cases, ensuring the handover between automation and human control is crucial.

Why Funders and Governments Must Act Now

Supporting applied research into mixed traffic AV safety is not optional — it is the only way to responsibly scale AVs. Funders have the opportunity to:

• Back cross-disciplinary research that blends traffic engineering, AI safety, and behavioral science.
• Invest in simulation environments to test AV-human interactions in realistic Indian and Australian contexts.
• Enable the development of regulatory frameworks that are proactive rather than reactive.

By doing so, governments and grant institutions can ensure that AVs deliver on their promise of safer mobility, instead of compounding the risks of already fragile road networks.