The Original Silicon Valley: Automation Didn’t Start in California

November 9, 2017 | Author: Joe Kenney

Uber’s self-driving car. A lot of people are talking about it, but if you think about it that technology started right here in the aerospace industry. Planes could “fly” themselves for years with the help of autopilot. Our traffic avoidance collision system (TCAS) and our enhanced ground proximity warning system (EGPWS) have been keeping planes from crashing into obstacles for years.

The truth is, despite being heavily regulated, the aerospace industry has been ahead of the innovation curve since we first took to the skies.

Today, we’re shaping the industry with advancements you often hear out of Silicon Valley, such as artificial intelligence (AI), speech recognition and machine learning. Now you might be saying to yourself “Joe, this technology already exists,” but the reality is when “Alexa” gives you the wrong answer the probability of anyone getting hurt is pretty low. Whereas, we cannot afford to provide pilots with inaccurate information. The integrity of the data we are supplying has to be spot on because we are talking about the operation of an aircraft and lives are at stake.

Honeywell has been carefully thinking about how humans are going to work this type of technology. Although most people are on the fence about flying on a pilotless plane, we believe that you have to be able to strike the perfect balance of pilot awareness and reducing pilot workload with automation tools. Relying on AI and machine learning will help reduce the number of aircraft incidents by giving pilots more support — think of it like adding another co-pilot to the cockpit.

So what exactly are these tools and how will they impact you? Well, here are just a few examples.

Speech Recognition: Say Goodbye to Language Barriers

Our advanced technology team is researching speech recognition tools and incorporating machine learning to reduce the probability of error and overcome language barriers.

As you probably know, air traffic control (ATC) is in constant communication with pilots to make sure they’re on track along their flight path. Talking over a radio or even the phone in a cockpit can be very noisy, way noisier than in a car, and it makes it difficult to accurately hear what someone is saying.

Right now, we are working on a robust speech translation tool that will help pilots overcome some of these obstacles by transcribing what the pilot is saying to ATC or vice versa onto a screen. Providing the information on a screen will allow pilots to refer to it when needed and simplify the takeoff and landing process because pilots will have a visual confirmation of the instructions given. In the long run, such technology will also help prevent hazardous landing run-ins, like when actor Harrison Ford landed on the wrong runway in February of this year.

As well, this tool will recognize different dialects or accents. Although English is the native language in aviation, someone’s dialect or accent may make it difficult to understand verbal instructions. Our speech recognition technology will be able to distinguish what people are saying, despite their dialect or accent, and transcribe it. This will save a pilot time and help eliminate confusion.

Mic’d Up: Mapping Out Noises in the Cockpit

Some other exciting exploratory work we’re doing with AI has to do with cockpit sounds. It involves installing a microphone into the cockpit, and understanding the different sounds every switch or button makes. This information is incredibly useful to investigators as they determine what has happened in a crash.

After an aircraft incident, regulatory partners, airlines and others listen to the flight recorder or black box to try to piece together the timeline of events. However, these recordings can often be incredibly difficult to understand, because of the background noise. With the help of AI and machine learning, we could play that recording and identify every button pressed and when it happened.

Think about what would occur if we could repeatedly review each sound we hear and combine it with the power of AI. The benefits would be twofold: We could better solve incident investigations, and immediately alert pilots if they push the wrong button.

By using isolated cockpit sounds as an investigative tool, we can better determine what may have gone wrong in the cockpit before and during an incident. The information also allows us to work faster and smarter, and make sure that these kinds of mistakes don’t happen again. For example, this solution could warn a pilot that they pressed the wrong button and allow them to course correct before anything goes wrong.

The applications for this are endless, which is what makes it so exciting. If we were to bring this technology into the driver’s seat of a car, we could better determine driver responsibility after a collision and expedite insurance claims. It’ll also help avoid future accidents by gathering data to better protect drivers around the world.

Light Detection and Ranging: Scanning the Air to Protect Aircraft

Finally, my team is also using a technology called light detection and ranging (LiDAR), which autonomous car companies are also using to detect and avoid surrounding vehicles and other obstructions. LiDAR uses a laser to scan and map out approaching shapes to determine their location. Why is this interesting? Well, we are using this laser to measure the condition of the air ahead of an aircraft. It essentially sees upcoming external environments, and then calculates how that situation could impact various aircraft systems. For example, if a plane were to fly by a volcano, the engine could ingest ash. Once ingested, the ash could transform into glass and cause significant damage.

The laser is able to study the approaching air quality, so a pilot can tell if he or she is flying into ash, ice or other detrimental conditions. Also, blending this LiDAR technology with our existing weather solutions such as Connected Weather Radar can help pilots make the safest decision based on surrounding weather patterns and help them steer clear of any dangerous or severe weather.

Powering the Future of Transportation

The technologies that we create at Honeywell are powering industries beyond aviation. Without Honeywell, things like virtual and augmented reality, commercial drones, and vertical takeoff and landing commuter planes wouldn’t be possible. Right now, we're developing technology that may eventually be used to keep you and your loved ones safe on the roads.

When we think about the future of automation, it’s about transforming the way people live, making it safer, more efficient and predictable, while also creating a seamless experience. We are making the impossible, possible — and there is nothing more exciting than that. The road to Silicon Valley runs through Phoenix, and Honeywell is at the center.

Joe Kenney

Joe Kenney

Chief Technology Officer, Honeywell Aerospace​.

Comments