Over the weekend I read William Langewiesche’s recent book Fly by Wire: The Geese, the Glide, the Miracle on the Hudson
In avionic application, the fly-by-wire approach establishes certain parameters that guide the actions of pilots – for example making it impossible to stall the airplane, or obtain more than 2.5Gs which could impact the integrity of the aircraft. Creating bounded ranges and cementing curbs allows for quick, decisive decision-making in times of distress. Bernand Ziegler – the former head of Airbus, and champion of the fly-by-wire approach explains, “we give you guarantees so you can react as fast as you want without having to worry about breaking the plane.”
In avionic application fly-by-wire is designed to increase airline safety. At its root it shows what machine-to-machine communication is capable of – if you are willing to entertain the notion that an airplane is just a series of isolated machines. So while the book narrowly focuses on fly-by-wire as it relates to aircraft control and flight safety, the applications for machine-to-machine communication, and a system involving pre-established parameter limits has much greater application.
In the end, anything involving a series of decisions (especially where safety is a concern) will eventually be influenced by a fly-by-wire-like approach. Fly-by-wire is progression of electronic systems that override potentially deadly user decisions wrought by a high stress situation. In this sense, we could increasingly see fly-by-wire influence things like automotive control. Automobiles are already a complex web of systems. I’ve talked in the past how sensors or MEMs are influencing personal technology. MEMs are ever-present in automobiles – from tire pressure gauges to proximity sensors in the bumper. Imagine taking the next step in connecting some of these sensor systems with the control mechanisms of the vehicle. This could include things like speed control when certain environments exist (weather, road surface, light) or break control to proximity sensors.
We can also think about machine-to-machine communication more broadly. A fly-by-wire approach that creates what I’ll refer to as “decision curbs.” Take for example mundane things like statistical modeling software. Statistical modeling software today requires a series of informed decisions. First a model must be specified followed by a series of statistical tests dependent on the results. At times, inexperienced analysts will attempt to do things that violate “rules.” Here a hierarchy of fly-by-wire-like inputs could override (or more efficiently guide) user direction. Autocorrect spelling might be analogous to fly-by-wire and provide a ready example of how systems will increasingly “self-correct” user control.
Today a fly-by-wire approach makes sense where high stress and limited time can influence decision making. Fly-by-wire will start by influencing high-value processes like jet travel, but the cost of taking a fly-by-wire approach will decline precipitous and soon be cost-effective to influence processes further down the value chain. Machine-to-machine communication will increase efficiency while lowering user input. An unintended consequence is that these type of approaches likely decrease the average quality of the user without lowering the quality of the output – certainly a concern voiced by pilot associations when it comes to core fly-by-wire systems.
Computer systems will increasingly step-in for menial tasks, periods of high workload, and times when a programmatic approach can be taken. This will allow for quick, flawless (or flaw-lowering) decision-making.
