NavList:

Let’s think about this idea of a single ‘backup’ for a moment.  For centuries there was no such thing as a 100% available, reliable, and accurate navigation system or anything like one.  Professional navigators, and they were mainly mariners, joined together what little information they could obtain from a multiplicity of sources, the lead line, the sea bottom, the sea colour, clouds, seabirds, smell, visual bearings, and celestial to make their best estimate of their position.  Not surprisingly, occasionally they were tragically in error.  

During the 20^th Century more reliable methods began to emerge, RDF, mechanical DR devices, GEE, Radar, Consol, Decca, and inertial,  but still they were by no means available World-wide.  Accuracy, whilst better, was by no means perfect, and all had their quirks and ‘gotchas’, so that the professional navigator was still balancing the output of one device against that of the others.  It was still a ‘check, compare, and decide’ situation.  A 1960s trainee air navigator flew his first two or three sorties using a manual air-plot even though the aircraft had a mechanical air position indicator (API).  Even after he was allowed to use the API, for the first 20 minutes of every sortie, he ran both a manual and a mechanical air-plot just to check the API was working properly.  Eventually, he was introduced to an aircraft with Doppler drift and groundspeed available feeding a mechanical ground position indicator (GPI).  The GPI with its various feeds allowed the navigator’s role to change to managing a system.  The system produced the aircraft’s position.  The navigator’s role was to monitor and manage the inputs, heading, airspeeds, and updating fixes based upon radar, RDF, visual, celestial, and if available, inertial.  It got so complicated that USAF’s B52 and the RAF’s V Bomber fleets employed two navigators per aircraft, and if they achieved positional accuracy with everything working within 200 metres on 50% of occasions it was considered very good indeed.  After three hours flying at 480kts at high level with all inputs switched off except celestial, three miles was considered quite good, and ten miles wouldn’t have prompted a major enquiry.

Then suddenly GPS arrived upon the scene, an aid that soon became so nearly 100% available, accurate, and reliable that its operating authority felt it necessary to get the satellites to lie slightly in their navigation message in order to deny its ultra high accuracy ‘on the fly’ to all but its own military and their friends (Selective Availability, which remains set to zero since 2000).  Moreover, the advent of large scale integration of microelectronics enabled position and speed to be calculated and navigation problems to be solved within a box initially little bigger than a match box and lately no bigger than a wrist watch.  On top of that, GNSS could only work by utilising nanosecond timing, which was quickly adapted as the master for commercial and financial timing and electrical engineering synchronisation world-wide.

Soon everyone with $100 to spare could become an instant navigator, and the world of commerce and engineering became tied to GNSS timing, encouraged perhaps by the erroneous belief initially that spread spectrum transmissions would be very hard to jam.  Well that’s what they were telling me in 2000!

So what do we do now?  As Frank has said more than once, the best ‘back-up’ to several satellites in a single GNSS system failing at once is to use one or more of the remaining satellite systems, possibly using some sort of automatic voting similar to RAIM.  In the unlikely event of all alternative GNSS becoming unavailable, then there is unlikely to be a universal ‘back-up’ solution, because users’ circumstances and requirements differ so much although the proponents of e-loran are trying hard to convince us that theirs is the best overall single solution.  

A yachtsman in mid ocean might well benefit from using celestial, even if he had to wait several days to see the Sun, Moon, planets or stars, but to take a topical example, what about a super tanker navigating the Strait of Hormuz.  You could expect to be in trouble even before the sextant could be brought to the bridge.  A nation’s financial services might seriously crash, and power grids could be thrown off line.  Celestial wouldn’t be much help here.  Any single back-up would need to be running in parallel in order to provide seamless changeover. 

The acronym PNT (Position, Navigation, & Timing) is a good one.  We need to consider all three before looking for solutions.  Solving the navigation problem is the easy bit.  Any simple microelectronic circuitry provided with a reasonably frequently updated database can do it can and already does do.  For position critical situations, such a system could provide position if provided with a number of continuously available or singly inserted inputs: depth, radar, visual, GNSS, e-loran, inertial, each with its own vote operating though a suitable system of filtering and artificial intelligence.  Unless someone comes up with an inexpensive star tracker which can see though several thousand feet of cloud, I can’t see celestial ever becoming a prime input to such a system except possibly in high flying aircraft.  Nanosecond timing might still be available from such a system, especially if it includes e-loran, but commerce and engineering might well find it easier to split from the position/navigation solution and look to ground based atomic clocks connected by ground line or space based ‘super clocks’ operating separately from GNSS satellites.  DaveP