Infosys launches aviation cloud it claims can halve lost luggage

Infosys has sent a digital transformation platform for commercial airlines down the runway and claims it could reduce lost luggage by fifty percent. The Bangalore-based IT services company's "Infosys Cobalt Airline Cloud" (ICAC) is said to “deliver personalized experiences, optimized operations, and net zero journeys for …

[Stu J]

Marketing BS

All of these types of systems have already existed for years, produced by companies whose entire existence is tied to their deep knowledge and expertise of the operational complexity of the commercial aviation industry.

I've no doubt that introducing this kind of software system at an airport delivers benefits. What stops it from happening at every airport - or makes it more difficult - is the cost of implementing and integrating with existing infrastructure, and training staff who just don't have time in their day allocated for training.

[jmch]

Re: Marketing BS

"... introducing this kind of software system at an airport delivers benefits..."

The software can only deliver benefits if the hardware and wetware allow it. There needs to be the physical infrastructure, and trained and capable human operators in place to deal with taking bags from point A to point B within X minutes. It's great if better software and interfaces (screens, tag scanners etc) can provide simple instructions for the optimal flow of each baggage item, but it also has to be implemented in physical space. The software also needs the capacity to be informed of where a bag *actually* is, rather than where the software has instructed that bag to go, and systems for retrieving misplaced items quickly and efficiently

[katrinab]

"Other tasks Infosys promised ICAC can tackle is turning legacy workloads into composable functional capabilities that reside in the cloud; improving crowd control with AI and optimizing route planning for decarbonization – an effort that coincidentally could also reduce flight costs."

So translated into English, it is a load of marketing buzzwords and the product doesn't actually do anything.

"turning legacy workloads into composable functional capabilities that reside in the cloud" - it runs old software on new hardware

"improving crowd control with AI" - crowd control is not a problem that luggage handling software needs to solve, therefore it does nothing

"optimizing route planning for decarbonization" - the optimal route for luggage is it to take the same plane as its human owner. Obviously that doesn't always happen, but taking decarbonization[sic] into account isn't going to help that. Therefore it does nothing.

[Doctor Syntax]

"Therefore it does nothing."

But it does do something. It transfers money to Infosys.

[jmch]

5%???

"reduce the number of containers used to load and transport cargo, known as Unit Load Devices (ULDs), and shrink the number that go missing to less than two percent through machine learning-enabled tracking methods. The current industry average of lost ULDs is five percent."

I know luggage going astray happens quite a lot, but 5%??? That on average for every plane with 200 passengers, 10 will end up without luggage?? That's far worse than I ever imagined!!

If I am correctly understanding what ULD refers to, these are huge stonking metal boxes, how does one of those 'go missing'??? I guess it rather means "we took it to the wrong place" and/or "there are thousands of these boxes in the luggage warehouse and we don't know which one is which"

[yoganmahew]

Re: 5%???

Generally it means the box (empty) is not in the place the ULD tracking system says it is. Like another poster says, wetware has to follow the process. If they don't scan the ULD as arrived or loaded on a plane (a problem when they ad-hoc change them due to knackerage or size limits), they end up wrong and hard to correct.

[Doctor Syntax]

Re: 5%???

The best ML can do there is learn that it's usually the bastards in shed 3 who CBA to get their dodgy scanner fixed. But everyone else probably knows that already.

[An_Old_Dog]

Re: 5%???

It's rather like how railroad cars are misrouted. I once saw a string of twelve flatbed railway cars, each one with a (military) tank chained to its deck, on a track siding in a tiny country village in the back end of nowhere. I'm pretty sure the locals were not of such a great threat that the army decided to deploy tanks against the populace there.

[Flicker]

Re: 5%???

Airports around Africa are littered with the things just outside the peri fence with people sleeping / living in them. 5% maybe an underestimate....

[hoola]

Eh.....

Forgive me for being cynical but just how can an "aviation cloud" stop luggage from being lost?

Stuff already has a barcode that is unique. One of my cases had an RFID tag stuck to it on one journey.

[Anonymous Coward]

Re: just how can an "aviation cloud" stop luggage from being lost?

It can't. END OF STORY

Does this system interface with ALL the different baggage handling systems that are in use around the world? Unless it does and can receive updates for every bag as it travels through the system and produce a dynamic map of every bag in transit then this is DOA.

The current systems mostly work because the system being used to track lost luggage is one that is indirectly owned by the largest airlines in the world.

Then... (and don't get me started) there is the security aspect of it all. Having it all in the cloud is a stalkers delight. Your journey could be tracked by a stalker or 'the man' through the journey that your bag takes.

[Stu J]

Re: Eh.....

Yeah the barcodes aren't unique, they get reused. Industry standard problem - 10 numeric digits, 1 for the bag type (usually '0' 95% of the time), 3 for the airline (e.g. 001 for AA, 125 for BA), leaving only 6 for unique bag identifier.

Large carriers get through 1,000,000 bags in a matter of days - which means if your bag hasn't made it to it's destination as planned by the time the number is reused, the chances of it ever getting back to you start diminishing ever further.

[Richard 12]

Re: Eh.....

So they just need to move to IPv6?

[An_Old_Dog]

Unique Serial Numbers

How much problem would it be to move to an IBM PS/2-style serial number, composed of alphanumeric characters, omitting 'B' (can be confused with '8'), 'I' (can be confused with '1'), 'Q' and 'O' (can be confused with 0), and 'Z' (can be confused with '2')?

[Stu J]

Re: Unique Serial Numbers

A massive problem. Bearing in mind that bag tag numbers are used for interline transfers between different flights between any pair of airports in the world.

Every single airline would have to update their backend software, and if necessary their mobile apps used by customers and crew/dispatchers.

Every single service provider would have to update their messaging and systems integration platforms.

Every single bag tag printer and self bag drop kiosk would have to be flashed with updated firmware (the printers as it stands can be told "print 3 tags starting with 001234567" and are 100% expecting something numeric)

Every single baggage handling system would have to be updated to handle new barcode types, in every single airport.

Every single baggage reconciliation and tracking system would have to be updated.

There's a buttload of stakeholders, a ton of testing, lots of dependencies and a massive dash of politics. And it would be a massive struggle to ever prove that there's a solid business case for getting the industry as a whole to agree to fund it. Not least because some airlines stand to benefit from it significantly more than others, but they'd all need to do it to maintain interoperability.

[An_Old_Dog]

Y2K-ish Problem

You're right about the extent of the problem, but it sounds a lot to me like, "Well, sure, we only coded two digits for the year, but everybody does it that way."

That everyone did it that way doesn't make it any less broken and stupid. And somehow, everyone copes when other major changes are made in, say, local tax rates. I'm sure there were/are some unimaginative programmers who coded the local tax rate as a numeric value, e.g., "PURCHASE_PRICE = (OUR_COST * 1.20) + (OUR_COST * 1.20) * 0.01;", but if you buy a badly-coded program or a badly-designed car, you know what yiou have to do about it.

[Magani]

Which is it?

"...claims it could reduce lost luggage by fifty percent."

So does is nick half the contents, or chop the lost bag in half?

[Bitsminer]

5 days

A certain American airline (whose initials are "American Airlines") held my bag in NYC for 5 days. They didn't release it because the staff at the baggage handling warehouse couldn't be bothered to move a bag to the correct truck for delivery.

Imagine 10,000 bags on the floor of a big warehouse. I suppose that could be disheartening, but that's the job. At least they scanned the bags daily. I got daily updates that it was "still" in NYC, "and your baggage delivery is very important to us" etc etc.

It took four days to get the correct phone number to the warehouse so I could talk to the staff and beg and plead. Only then did they take pity and load it onto a plane.

No amount of software is going to make people do their jobs, that's what supervisors are for. Ain't no supervisors in the cloud.

[joed]

Re: 5 days

No supervisors but hypervisors aplenty.

[stiine]

How hard can it be to put a RFID reader on the loading door and not let the airplane back away from the terminal if 1) luggage that should be onboard isn't or 2) luggage is onboard that isn't supposed to be.

I'd suggest that if start delaying flights for improperly handled luggage, the airlines would have the problem fixed in less than a day.

[Anonymous Coward]

I was thinking a barcode scanner for the folks loading the plane - if it's the wrong plane, the scanner shakes and flashes red. (No sound; they're outside on the tarmac where it's LOUD.) If loading is done but some luggage isn't on the plane, then prevent takeoff for at least ___ minutes while attempting to locate the luggage.

Combine that with an automatic payment to anyone whose luggage is lost. $200 minimum, with the ability to make the airline pay the real value of the luggage and contents. Enforce this by law, not have it be optional for the airlines. Customer returns HALF the money if the airline finds and returns the bag. I once had them lose my suitcase with a $700 suit; they offered to pay me so little that the suitcase itself, without the suit, was worth more. (Thankfully they found it and got it to me in a couple of days.)

[Stu J]

That already happens at most major airports. They even get told not to load bags if you've not yet cleared security. For containerised holds (i.e. anything bigger than a 737, or any Airbus) they get told which container to load it into to optimise transfers. None of this is new.

Generally speaking if a bag arrives on the right baggage makeup belt (think of a baggage reclaim equivalent temporarily dedicated for a specific flight where baggage handlers take bags and put them onto a cart or into a ULD container) there's a very high chance it ends up on the right aircraft.

Problems happen if the tags fall off, or don't get read properly, or if the software integration from the airline host systems to the baggage handling system in each airport falls over and stuff gets dumped for manual sortation instead of onto the correct makeup belt. Or the bag hasn't been carefully loaded on the belt and gets snagged and falls off.

The other problem is if there's a delay unloading a late-arriving connecting flight - generally speaking it's orders of magnitude cheaper to let connecting passengers make the connection and let it depart on time without their bags, than it is to delay the connection for everyone and cause a domino effect of delays.

[Stu J]

Most bag tags still aren't RFID-capable (unless you're flying with an airline that has spent 10s of millions and gone all-in on RFID - like Delta). They're old-school barcodes that have to be manually scanned.

Also the cost of delaying an aircraft so that one missing bag makes it is so high that it's unrealistic to expect perfection. Shit happens, bags fall off conveyor belts, tags that weren't put on properly fall off, equipment malfunctions, software malfunctions.

Repatriating a couple of bags and providing compensation is expensive, but not as expensive as delaying hundreds of passengers can be for 20-30 minutes while you find a missing bag - that delay can lead to missed connections (which increases likelihood of mishandled bags), people having to be put up in hotels overnight, aircraft getting later and later throughout the day...

Trust me if airlines could wave a magic wand and get perfection, they would. The reality is they could spend all their profit every single year on it and still not get close. Think of it like a giant logistics operation like UPS or FedEx run, except you're additionally having to try to co-ordinate the route the package takes with a person on a journey, and unlike UPS or FedEx you don't own the infrastructure at most of the airports you fly into. A lot of the times it works fine, but sometimes things go wrong. There's too much wetware involved,

[Fruit and Nutcase]

Hi Dad

Rishi is for ever worring about luggage going missing at airports. I remind him that we don't take commercial flights so it is not a problem for us. Still, is there anything that you can do to reassure him? As air travel is the only form of transport that he can relate to with the voting public, it would be a great help

love

A