Flight Management Systems for Singles and Light Twins
Stephen B. Miller
Re-printed with permission from the Aug. '97 issue of Avionics News magazine.
Flight Management Systems (FMS), while normally found only on larger
aircraft (i.e., bizjets and up), have a potentially unique application for
singles/light twins, largely due to the availability of small "embedded
processor" systems. Historically, the FMS is used in a
"multi-engine/multi-crew" environment, whereas single-engine aircraft are
virtually always operated by a single pilot and light twins are frequently
operated in the same way. This brings up some interesting "what-if's". For
example, what if an aircraft of this class had an FMS coupled to the
autopilot (including all the usual nav sources) and could communicate via
the MODE S transponder? What if this system also possessed a relatively
high-level of software such that all the necessary "human-like" decisions
could then be made? The answer to these questions is that you'd have a
full-time "virtual" co-pilot, capable of unassisted flight from A to B,
ENGINE FAILURE: In a "single" operating enroute in IMC, getting the
aircraft down "dead stick" is no mean feat, to say the least. The
multi-tasking capability of a system such as this would enable the aircraft
to quickly set up the best available approach at the best airport within
gliding range and proceed to execute it on a "how-goes-it" basis, with
constant updates for both altitude and direction. Communications would be
handled simultaneously through the MODE S data link. All decisions and
current status could be visually displayed to the pilot in real-time. The
FMS would position the aircraft so that a manual landing could be
accomplished at the appropriate time.
PILOT INCAPACITATION (single pilot): In this case, particularly in
IMC, passenger survival is highly unlikely without the sort of back-up the
FMS could provide. This emergency has a distinctly different "wrinkle" from
the previous one, in that both power and fuel are available, allowing the
aircraft to select the best available precision approach at the airport
having the best weather as well. In this case, however, "auto-land"
capability is also required (more about that later).
THE DESIGN CHALLENGE: In order to provide these capabilities, the FMS needs
to sense/interpret all of the inputs available to the pilot and to respond
with "pilot-like" skill and intelligence. This would obviously include such
things as frequency/autopilot mode selection and the ability to control the
engine and all other "ancillary" devices (gear, flaps etc). Providing the
necessary "pilot-like" decision making capability is by far the most
difficult requirement, but is feasible through extensive software
development. The level of "knowledge-based" software required dictates that
it be developed by people who are also experienced pilots. These folks
exist, but they don't exactly grow on trees. Can you envision anyone else
attempting to write the code to manage a dead-stick instrument approach,
for example? The real difference between an existing FMS and this version
is the combination of autopilot coupling, total control of the aircraft and
the high level of integration software, the goal being to greatly enhance
both the safety and utility of singles and light twins.
HARDWARE CONFIGURATION: Many system configurations are possible, though in
a mature design there is no good reason to separate the
autopilot/controller functions. It makes more sense for the controller to
perform all autopilot tasks, issuing servo commands based on
instrument/pilot inputs over a common bus. For a "conceptual" prototype
design, however, the simplest initial approach is to envision the system in
terms of modifications of existing hardware wherever possible. Viewed in
this light, we have the following components:
- SYSTEM CONTROLLER: This is the "embedded processor" mentioned
previously; think of it as the "brains" behind the operation. This unit
also has sufficient I/O capability to communicate with all other system
components using a standard interface, such as RS232, RS449, one of the
ARINC busses etc. It is beyond the scope of this article to deal with the
details of particular system architectures. Communications can be
implemented through individual RS232 ports from the controller to each
system component or through a common parallel bus, for example. The
controller gathers and analyzes the data from the other system components
and issues the necessary commands. The point is, there are a number of
these OEM processor systems available, readily adaptable to this task,
either in present form or with packaging modifications, for example. They
generally are rack-mounted or simply "stacked" PC boards, ranging in size
from extremely compact to "a little larger than I'd like". Many of these
will mount remotely behind the baggage compartment or in the nose of a
typical light twin, for example.
- AUTOPILOT "SYSTEM": This refers to the autopilot and all
associated instrumentation/servos, with the following modifications:
- The autopilot needs a communications interface as
explained in the previous paragraph in order to "talk" to the controller,
which must be able to control the autopilot in exactly the same fashion as
the pilot can.
- The FMS must also be able to control the engine, gear,
flaps etc as mentioned previously in order to handle the job in the case of
pilot incapacitation. For ease of visualization, consider a typical
"Piper-style" control quadrant incorporating gear and flaps as well. This
can be viewed as a single "servoized" package, representing an additional
autopilot component, although a complex one. Actually, if autopilot
redesign is to be avoided, this "component" would be controlled directly by
the system controller, operating in conjunction with the autopilot, until
such time as the autopilot and controller functions are eventually merged,
as probably would be the case in future designs.
- NAV SOURCES: In addition to the usual "steering" inputs to the
autopilot, the controller would also communicate with these components for
status/control (mode/freq. selection etc). Again, the controller's job is
to "emulate" the pilot. Whereas GPS receivers would typically provide, say,
an RS232 port, older VHF NAV sources would not, leaving them out of the
picture. Considering the rapidly accelerating dominance of GPS, this is not
likely to present a problem in the long run.
- TRANSPONDER: Obviously, this would be the controller's only
means of air-ground communication using the MODE S data link. Similar
coupling to a TCAS unit would likewise be a definite "plus".
THE CHALLENGE OF AUTOLAND: This capability has never existed on singles or
light twins. The development costs/technical problems have always exceeded
any perceived need. In the case of the FMS described herein, however, the
need is real enough if the aircraft is to be "smart" enough to extricate
the passengers from certain disaster should the pilot become incapacitated.
I don't think anyone would argue that controlling a small aircraft near the
ground in a typical turbulent crosswind is a far cry from doing the same
thing in a '747. The problems of precise, instantaneous lateral (DGPS?) and
vertical (RADIO ALTIMETER?) guidance/control are substantial, to say the
least. No doubt, other fast-reaction sensors such as accelerometers would
also be needed. Controller response times would also have to be fast enough
to take care of sudden, unexpected changes in flight path (i.e., the
vertical gust that leaves the aircraft "high and dry", for example). All of
of this would have to be combined into a system whose reactions could not
be discerned from those of a human pilot, a daunting challenge if ever
there was one.
SYSTEM POWER: After all this effort to develop a system which is "always
ready" to provide whatever assistance is required, power for its operation
must be assured, even on a single with a failed engine/alternator. This
means a redundant power bus must exist, invoked automatically when needed.
The backup power source could be a high-energy battery, air-driven
alternator or other means, but must be free from "contamination" by other
system failures. This aspect of the design is not trivial and must be
thoroughly thought out if it is to be commensurate with the system goals.
THE MARKET, WILL IT EVER HAPPEN? It depends. If a viable manufacturer feels
the market is ready and that this system could be made "affordable" (can
anyone really define this?), the attempt will probably be made. It could
also fit in well with NASA's efforts to create safer aircraft.
Incidentally, my guess is that only an existing autopilot manufacturer
could even begin to think of a development such as this, so the potential
field of players is limited right from the beginning. If this system were
available today at the right price, how big do you think the retrofit
market would be? How about the new market, now that the industry appears to
be "on the mend"? Obviously, there are more questions than answers here,
but I think you get the picture.
This overview has been presented in an effort to point out ways in which
existing technology might be channelled to provide greater utility/safety
for small, high-performance aircraft. Hopefully, a sufficient market exists
which will make these things possible in the not-too-distant future.