The sheer amount of "snake oil" bolt-on products sold to fleets in the 90s and 00s means the industry now demands rigorous third-party validation. The best open source testing I've seen comes from Mesilla Valley Transportation Services (https://www.m-v-t-s.com/certified-technologies/). They don't rely on ECM dashboard readouts or pump receipts. They run a control truck and test truck simultaneously on a track or controlled highway loop so weather, air density, and wind are identical. Anemometers on the trucks algorithmically factor out wind speed/direction, tyre temp, and weight. Tests are designed to answer the actual question rather than tick a standard's boxes. Their chief test engineer was ex NASCAR, IndyCar and F1.

My team actually tested a dual-fuel system with MVTS. The engineering was sound, testing proved a slight thermal efficiency improvement, but cheap US diesel versus the cost of the alternative fuel meant the ROI was non-existent. The physics worked, the math killed the project. (The economics work in Europe, thankfully.)

A lot of what's being discussed in this thread is already standard too. Mirrorless cabs are more the rule than the exception now (as pjc50 linked above), close-coupled trailers with gap fairings, wheel covers, under-trailer aero, all commonplace. A couple of current ones worth knowing about: the Scanias with the "Super" powertrain run a really aggressive overdrive at around 900rpm at the 90km/h (56mph)EU limit, where a conventional truck sits around 1200rpm. Requires huge low-end torque but claims 5-8%. The Mercedes Actros L claims 3% from its slippery front end alone. It is absolutely hideous though, so make of that trade-off what you will.

The biggest factor though is driver variability. All the aero and powertrain hardware in the world doesn't matter if the driver thinks they are the stig. The biggest shift in the last decade has been removing human inconsistency from fuel economy entirely.

Automated Manual Transmissions are now completely dominant in modern fleets. They shift perfectly for fuel efficiency every single time, totally capping the penalty of a bad driver. Layer on top of that Predictive Cruise Control. Modern trucks use GPS and 3D topographical maps to "see" miles ahead. The truck's computer knows exactly when to back off the throttle before cresting a hill to coast over the top, and exactly how much momentum to carry into the next dip. A human driver relying on their eyes simply cannot compete with a powertrain that knows the exact gradient of the road three miles ahead.

https://www.volvotrucks.co.uk/en-gb/news/insights/articles/2... : removing mirrors allows for much smoother cab airflow

https://go2stream.com/blog/aerodynamic-truck-legislation-rea... : UK legalization of fishtail-like devices

https://www.kudauk.ltd.uk/aerodynamics-explained : Kuda on the UK allowing higher loads, and therefore benefiting from extra wedge devices on the top of the cab.

I'm sure there's a lot more out there. The eventual switch to electric will probably come with another round of aerodynamic improvements to maximise range, as with cars.

There are other things to consider as well when building a truck, as mentioned the maximum length limits, comfort, power price and more.

https://www.colani.org/luigi_colani_Product_design_museum/Tr...

Makes you wonder if/why truck companies don't have aerodynamics experts on their payroll.

https://www.ccjdigital.com/regulations/article/15291029/acha...

This is the most recent news I can find about them, though, so unclear if these engines will ever be produced for road vehicles:

https://www.govconwire.com/articles/ga-asi-david-alexander-a...

See also: https://achatespower.com/resources/

I was hoping we'd see these engines running generators in Edison trucks one day.