My personal perspective in automotive technology is dominated by sustainability and minimalism, but I also have some professional interest and therefore will dig a bit more deeply.
Autonomous vehicles (AVs) (specifically land vehicles) are something for which I’m very hopeful since they could reduce or avoid the need for ownership and similarly could reduce the number of total vehicles needed to exist given that the capacity could be defined in terms of the amount of contemporaneous volume of travel rather than what translates to what is closer to a capacity that effectively represents everyone traveling simultaneously (where this feels like a generally recurring fallacy that likely has a name). This also lends itsely to further optimizations in terms of both reducing travel time (i.e. reducing traffic) and introducing queuing with corresponding incentives to reduce needed capacity by absorbing bursts (similar to existing peak/congestion pricing models where a given traveler pays extra or waits some amount of time). Unfortunately In addition to many of the immediate practical challenges, I’m also somewhat skeptical in the realization of such reductions in that it may challenge the business models of many involved players (i.e. manufacturers) and the persistent risk that such gains in efficiency are offset by rebound effects.
At present there remains a fairly wide gap between the current state of artifical initelligence (AI) systems and their envisioned end state and their advertised abilities (which should be covered elsewhere) and AVs are a particularly clear example of this given the surrounding safety concerns. AVs may operate in environments with a virtually unlimited number of variables which can fall very easily into the cracks of modern AI. In particularly chaotic environments such as India, stocastism may be embraced{atime=“2024-07-01}, but the more typical approach by US companies seems to be seeking to cover the fantastically long tail. Operating in environments where such variables can be controlled is a wise and employed tactic, though it is important in other cases to make sure that variations are not aggregated away when presenting statistics and that instead significant conditional probabilies are expressed. This can help paint an appropriately nuanced picture{atime=”2024-06-29} which may also feed more directly into questions around how the levels of autonomy within the system could safely integrate with propsective human intervevention when necessary (where this may also be more common than perceived ). This translates to not not only prospective deficiencies within the systems but risks around dangerous fallback behaviors.
Beyond the AI concerns, there will also be ongoing work in compensating for prospective failures or deficiencies with sensors, so this is a particularly hairy problem from multiple angles.
Electric vehicles are a place where China is poised to claim a fairly dominant position. In addition to attempting to produce competitive products, economic strategies such as tariffs along with more cut-and-dried bans may also be applied to at least relieve some pressue for domestic copmanies.
Vehicles connecting to a range of infrastructure and other devices is gaining traction (maybe pun?). There have been some initial projects involving smart highways and mass transit adjusting stop lights along with recommendations from the department of transportation.