A Call For Dust-Off

The Search For A Reliable ELT Replacement

When Kenya Airways Flt KQ507 went down shortly after takeoff from Douala Cameroon, the search started 250 kms away. Even though Douala ATC had declared the aircraft incommunicado when it didn't check in passing 5000ft in the climb, they were not to know that it wasn't still airborne.

Alex Bayeck, an official with the Cameroon government, initially announced that the Nairobi-bound plane had gone down 250kms from Douala, the commercial capital of Cameroon. The search moved closer with each subsequent announcement, but initially to an area near the town of Lolodorf, about 140kms (90 miles) southeast of Douala.

It wasn't until more than 40 hours had elapsed, that the search was redirected back towards Douala. The impact crater was finally found among mangroves just over 5kms from the runway's end.

The feedback that ATC got from the French Satellite Tracking Station in Toulouse was taken at face value. Once the satellite data is in, it takes less than a minute to forward the received Emergency Locator Transmitters (ELT) fixing data to any signatory nation. Data offered by Toulouse was misleading because it centered the initial search area 150kms away, along the 737's planned track.

Communication Minister Ebenezer Njoh Moulle told reporters in Yaounde: "The information they furnished pointed to two areas, one in South Africa and the other in Nyong and Soo (southern Cameroon). That is why the initial search for the plane was directed to Lolodorf and its surroundings, which is about 150 km from the actual crash site. The question is why the plane's distress signal frequency failed to operate automatically, as ought to be the case." Why, indeed?

The reason is because of the specs for the ELT. TSO-C91a, EMERGENCY LOCATOR TRANSMITTER (ELT) EQUIPMENT prescribes the minimum performance standard for ELTs operating on 121.5 and 243.0 Megahertz, while TSO-C126 - 406 MHz EMERGENCY ELT prescribes that for a 406 MHz ELT. The FAA urges turbojet-powered aircraft operators to install an ELT that operates on 406 mhz.

In addition to the 406 MHz ELT's stronger signal and the almost instantaneous detection by geostationary satellites, the 406 MHz ELT signal can be coded with the owner's identification or aircraft coding. This ID permits Search and Rescue Coordination Centers to contact the registered owner by phone and verify if the aircraft is flying, missing or safely on the ground.

This capability permits a rapid SAR response or allows the operator to deactivate a 406 MHz ELT that is inadvertently transmitting.

In addition to its many other benefits, newer 406 MHz ELTs have been designed with the capability to transmit an aircraft's last known position. This embellishment further reduces the 406 MHz ELT's already small search area.

The weather satellites that carry the SARSAT receivers are in "ball of yarn" orbits, inclined at 99 degrees. The longest period that all satellites can be out of line-of-sight of a beacon is about two hours. Some geosynchronous satellites have beacon receivers.

Since December 2003, there have been four such geostationary satellites (GEOSAR) that cover more than 80% of the surface of the earth. As with all geosynchronous satellites, they are located above the equator.

The GEOSAR satellites do not cover the polar caps. Since they see the Earth as a whole, they see the beacon immediately, but have no motion, and thus no doppler frequency shift by which to locate it.

However, if the beacon transmits valid GPS data, the geosynchronous satellites give searchers an instantaneous response. But why didn't this work for KQ507?

The explanation would seem to be in three parts. First, the "g"-activated ELT located in the tail of the aircraft would only have emitted a few short bursts before being immersed by the rapidly filling crater. In the short interim before the crater filled with water, that 406MHZ transmission may have only been picked up by two satellites within Line of Sight of the ELT in its location below the partly shielding crater sides.

The brief transmission would produce two Line-of-Position (LOP) arcs that intersect twice at very shallow angles. Visualize two intersecting circles overlapping at half radius. The two intersections are about 1.25 times the radius apart (i.e., many 100's of kms).

If a third satellite can't produce another confirmatory LOP, the two points of intersection will represent two rough fixes, only one of which will be approximately correct (and therefore chosen).

A four-satellite fix, by contrast, would be quite accurate and without any fixing ambiguity, especially after a brief resolution process based on continuing transmissions. (See an explanatory image at tinyurl.com/2ru99f.) ELTs below the lips of craters have done this before. However, a crash crater that fills rapidly with water provides a new challenge for ELT designers.

Perhaps the answer is a floating antenna buoy that detaches upon impact? Maritime Emergency Position-Indicating Radio Beacons (EPIRBS), carried by vessels and deployed either manually or automatically, are quite sophisticated and float, but aren't designed to take the impact of an aircraft crash. And of course any stored GPS position transmitted by KQ507's ELT may have itself been errored by poor GPS satellite reception.

The continuous transmission mode of 121.5 MHz beacons limits the capacity of that system. In addition, other users in that band also limit the capacity. 406 MHz beacons only transmit a 1/2 second message every 50 seconds (100 seconds for a swamp's crash crater to fill?). This, combined with the random time sharing of the spectrum and frequency spreading, allows receipt of hundreds of distress signals within view of the satellite.

Two satellite passes are necessary to resolve the fix ambiguity for 121.5 MHz beacons. However, the ambiguity for 406 MHz beacons can be resolved in one pass. This is due to the improved oscillator stability which allows the COSPAS-SARSAT system (tinyurl.com/2uwsxw and 2tfn2l) to determine the real location 95% of the time, and due to the ability to determine what type of beacon is activated.

This can save up to two hours in responding to the alert. The normal fix waiting time for 121.5 MHz beacons is 45 to 90 minutes (this is how long it takes on average for a satellite to come overhead).

Although this figure is the same for 406 MHz beacons when waiting for polar orbiting satellites, the waiting time is significantly reduced with reception by satellites in geo- stationary orbits.

The waiting time is typically less than 10 minutes. If the 406 MHz beacon has a valid GPS output, the accuracy is improved to 100 meters, which translates into a search area of less than 1 square kilometer.

Before looking at a possible solution, is there a requirement to resolve ELT vagaries and reliability, or was KQ507 just a "one off"? A similar protracted confusion surrounded the fate of Bellview Airlines Flt 210 that crashed on departure during a storm in a remote area just 20kms north of Lagos on Oct. 22, 2005. The impact crater of that Nigerian 737 in swampland had in similar fashion filled with water, immersed and killed off the tail-mounted ELT's signal.

However, in that case, no signal had been received so it may well have been destroyed by the impact. The sound of its arrival had been merged with the sounds of the thunderstorm in the minds of the local inhabitants.

On April 20, a Beech 18 twin-engined floatplane of Vancouver Island Air capsized on takeoff in the freezing waters of Jackson's Bay, British Columbia. The nine-passenger plane sank within 45 seconds, but the survivors held onto broken pieces of pontoon which were also taking on water and going down. Fortunately, a logger on shore heard the crash, jumped in an aluminum boat and used a shovel to paddle out to the seven survivors.

A Cormorant SAR helicopter arrived shortly after, but no thanks to the aircraft's ELT. The float-planes have ELTs located in the aft section but they won't function under water. The company now plans to outfit all its aircraft with portable ELTs that are water resistant and can float.

The portable transmitters will be placed by the exit doors for exiting passengers to grab (if indeed they remember to do that in their haste to leave and take dinghies/life-vests with them).

In the case of the Adam Air 737 crash on New Year's Day 2007, it took weeks for survey ships to detect the signal from the flight data recorder's transponder beacon sitting at a depth of around 1200m. That quickly faded and now a position accurate only to a few kilometers is held for that crash-site and its unrecovered recorders.

If it had been a controlled ditching, there may have been survivors for whom an ELT would've been vital, particularly if no life raft with an EPIRB had deployed.

From Jan. 1, 2005, all "aeroplanes" operating as "commercial air transport" on "long-range over-water flights" and (as defined in ICAO Annex 6, Part I) shall be equipped with two ELTs, one of which is automatic. These ELTs must transmit on 406 MHz and 121.5 MHz. The situation, viewed nation by nation, is unclear.

Some require portable (life-raft) ELTs so that rescuers can home in on survivors rather than the wreck, and some do not (tinyurl.com/3cr8rf). The question of flotation arises only inasmuch as it's assumed that an automated ELT will activate on land and a portable one will be carried in a liferaft. Water impact should be factored in.

In summation, there appears to be a number of problems with ELTs. First, they can be destroyed by impact; second, they won't transmit a signal when immersed; and third, having no redundancy and only the fixed aircraft antenna (or tethered for a deployable ELT), they may not operate.

Even if they don't fail, they may be below the lip of a crater or shielded by a cliff-face or ravine and unable to "see" a line of sight to the monitoring geostationary or polar- orbiting satellites. A search and rescue effort that is directed to the wrong area can be very unproductive and frustrating, particularly if there are survivors.

Is there another solution to these deficiencies, perhaps one that would restore vital reliability?

A company called Dust Networks produces a wireless network called "Smart Dust", a very modern signal networking and modularity concept (with a somewhat unfortunate name) that was articulated, publicized and developed toward commercial reality over the last 10 years by Prof. Kris Pister at the University of California Berkeley.

Like a Radio Frequency Identification (RFID) chip, Smart Dust consists of a large number (up to millions plus) of self-powered or solar-powered intelligent radio transmitters/transponders ("motes") that can collaborate with each other as a network or work alone, each one not much bigger than a flyspeck (image at tinyurl.com/2z9fgw). Released or ejected rearwards as a cloud, a partial population of "motes" will survive any sort of accident with high probability, then distribute locally on the breeze, settle on ground or water and be readable by satellite, portable, and surface equipment. See links page at tinyurl.com/2g4aza

Thus far, Dust Networks hasn't thought of an air rescue application for its technology. If the company does go for it, it'd be a nice gesture to call the system "Dust-off", in honor of those pioneering helicopter crews who homed on the smoke and dust of battle and extracted the wounded under fire in Vietnam.