Clearing the roadblocks on the path to full autonomous driving

Progress towards ambitious goals is rarely linear. The effort to achieve fully autonomous driving is no exception. It may disappoint those that had anticipated a “self-driving-car revolution” to learn that getting there will be more of a multi-year, step-by-step, feature-by-feature evolution. The approach adopted by automakers and regulatory authorities to safeguard lives and bolster trust in the technology: start small, authorize a limited number of features in narrowly defined conditions, and grow the scope of the technology from there. Cautiously. Incrementally. Safely.

That said, it would be misplaced to call the evolution slow. OEMs today are already selling models boasting Level 2 autonomy features that control driving speed, keep a safe distance from the car ahead, stay in their lanes, and offer countless other forms of assistance, from parking and overtaking to blind-spot and “backing up” alerts. The condition, however, is that drivers remain focused behind the wheel and prepared for the vehicle to turn off support at a moment’s notice.

The next stop, already visible on the horizon, is Level 3 autonomy. While it’s just another increment on the roadmap, in terms of the demands it places on the vehicle, it’s a huge leap. Level 3 requires the vehicles to autonomously authorize self-driving mode when permitted and puts the autopilot in charge. Meanwhile, drivers get a chance to kick back, enjoy a movie, or focus fully on a phone call, at least until the vehicle exits the authorized self-driving zone or loses confidence in data coming in from its sensors. Achieving the context-awareness this requires in a broad range of environments will be a tipping point on the road to higher levels of autonomy.

But the journey to Level 5 – no steering wheel, with no geographical restrictions – is fraught with technological, practical, and regulatory bottlenecks. These are some of the most challenging ones we see on the horizon.