Careers in AI — AI in Transportation
The transportation sector is becoming a data driven industry with features and applications far beyond Tesla and Google’s driverless cars.
Author: Andre Frade
What is the transportation industry about?
The transportation industry refers to the economy segment that provides infrastructure and means for the movement of people and freight (good, resources, materials, etc). The sector branches out according to the services and space — land, water, air — that companies focus on.
Why consider machine learning in transportation?
The transportation sector results from the collection of many individual moving parts that interconnect and influence each other.
Nowadays each moving part — a car, a plane, a ship — is often equipped with tens to hundreds of sensors that constantly gather data from their surroundings. It is also standard practice for most vehicles to be part of a cloud network where data from other vehicles is readily accessible, powering advanced and improved safety features for all the members of the network.
The amount of new data available every second is so vast and complex, that only advanced analytical tools are able to handle it fast enough in order to produce useful insights in real time.
The intersection
In the transportation sector, vast amounts of real time data are combined with machine learning technologies to address two primary aims. Some tools are dedicated to the improvement of operations, whilst others are used to reduce risk.
In general terms, operations may be improved using algorithms that address fuel efficiency, route and schedule optimisation, traffic management, staff management and pricing strategies, whilst risk reduction is often achieved via predictive maintenance and fraud detection. Either way, achieving the same goal with a private car, a commercial airplane or a cargo ship may entail different considerations, and AI tools must be adapted to each context.
In this article we will focus on the automotive division of the transportation sector. For a broader overview, check out Artificial Intelligence in transport.
The future
The use of sensory data, connection technology, and machine learning analytics is becoming an industry standard for the transportation sector. AI is predicted to be deployed even more widely and help to accelerate the maturity of autonomous vehicles. Vehicles are expected to be able to better communicate with each other and its surroundings, increasing the efficiency and safety of the transportation space as a whole.
Types of Applications
In this article we take a deep dive into the applications of AI to the automotive industry, as part of the transportation sector. Below are five of the most popular applications, which include 1) sensors and network technology services, 2) driving features, 3) driver monitoring, 4) car insurance and 5) car manufacturing.
1. Sensors and Network Technology Services
Cars gather data from their sensors, which they can use directly or upload and integrate with data from other cars. This data enables driver features such as driver assist and autonomous driving, but it also plays an important role in predictive maintenance and personalised marketing.
For predictive maintenance, connected cars use AI cloud services to anticipate problems and faults in the vehicle before any of its operations are compromised. AI tools constantly monitor the condition of the car and compares it against previously collected data to identify any situations where the car needs to be checked. AI cloud services can also provide personalised marketing. The car learns the driver profile and routine to produce relevant alerts each time the vehicle is in the proximity of a business that can serve their needs. That may be in a low fuel situation or whether a restaurant with the driver’s favourite cuisine is just around the corner.
2. Driving Features
Artificial Intelligence technology integrate sensor and cloud data to power advanced safety features, offering two levels of driving functionality: driver assist and fully autonomous mode.
In the driver assist mode, data is used to augment human capabilities and anticipate danger. Some of the features may include blind spot monitoring, lane and obstacle detectors, driver-assist steering, and emergency braking — all designed to help to reduce risk on the road and avoid accidents. In the fully autonomous mode, the driver hands over all the driving operations to the vehicle. These cars are equipped with state-of-the-art technology that makes them able to safely navigate the roads with full autonomy and machine precision. Many automakers and start-ups are working on AI applications for the automotive industry, but two companies take the lead for the development of autonomous cars: Google and Tesla.
3. Driver Monitoring
Some automotive machine learning software are designed to identify, recognise and monitor the driver, in order to prepare the car to their preferences and increase the safety of the trip.
Identification features detect whether there is a driver inside the vehicle. Recognition features are able to detect who is operating the vehicle, and adjust the car to their preferences — seat, mirrors, temperature, etc — to ensure a pleasant and safe drive. Finally, monitoring features are activated during the journey and constantly look out for the driver’s posture and any signs of fatigue or distraction, alerting them to refocus their attention on the road, if necessary.
4. Car Insurance
AI has found applications in the car insurance space, whether to provide risk assessment to drivers or assist on the accurate and fast completion of claims, when accidents occur. AI powered risk assessment integrate the driver’s personal data far beyond their driving history to accurately evaluate the risk and find details on how their safety might be compromised on the road. Despite the improved safety, when accidents do occur the car is able to provide on screen instructions on how to properly record details for the insurance claim. Additionally, AI is also able to inform the driver how to have the vehicle repaired and what will be covered by insurance.
5. Car Manufacturing
AI is not only powering new car features, but also taking over the production line and innovating the manufacturing process. Automotive plants include AI powered robots built to increase safety and efficiency in the workplace. Wearable robots help workers protect their knees, back and neck whilst giving them the mobility and strength to perform their tasks. Automated Guided Vehicles move materials around warehouses safely and without human intervention. Technical robots paint and weld autonomously, but their AI features empower them to identify defects or irregularities in materials and alert quality control personnel.
Types of companies
The transportation sector branches out according to the services and space — land, water, air — that companies focus on. Major companies are able to develop their own AI features in house, whist others acquire the capabilities from dedicated start-ups.
Examples of companies operating in the transportation space include:
- Land: Google, Tesla, Jnction, Oxplus
- Water: Stena Line, Olesinski, Ladar Ltd, Orca AI
- Air: Easyjet, Ryanair, Delta, Emtjets
Types of Jobs
Senior Machine Learning Engineer, Autonomous Driving
A Senior Machine Learning Engineer is expected to build new solutions to problems. Knowledge of Deep and/or Reinforcement learning and experience working with big data are often required if the projects are within the self-driving space.
Autonomous Vehicle, Security Engineer
An Autonomous Vehicle Security Engineer is responsible for ensuring the highest level of security of autonomous vehicles by engaging with all stages of feature and software development to identify any security issues in the product.
Software Engineer, Behaviour Prediction
A Software Engineer working on behaviour prediction is responsible to build the brain of an autonomous vehicle, such that it perceives the world around it and makes the right decision in every situation, guaranteeing that passengers reach their destination safely.
Product Manager, Simulation
A Product Manager is expected to work closely with a whole team of Data Scientists and Engineers to align the development of the product with the company mission. In simulation, the product manager ensures that the autonomous vehicles technology is properly tested.
Example of Interview Questions
Examples of Non-Technical questions:
Why are you interested in the company?
What makes you think you’re a good fit?
When have you failed and how did you learn from it?
Tell us about a time when you had to solve a problem with little to no information about it.
If I asked you to quit your job and start working with us tomorrow, which three things would your current manager miss most about you?
Rate yourself vs. your peers on a scale from 1 to 10.
Examples of Technical questions:
A fair 6-sided die is rolled twice. What is the probability of getting 2 on the first roll and not getting 4 on the second roll?
There are 100 products and 25 of them are bad. What is the confidence interval?
Reverse a string but ignore special characters.
Reverse a linked list in-place recursively and return the new head pointer.
Given a polynomial function with n terms and k degrees, how many partial derivatives can you form?
Write a command-line program to evaluate a set of equations.
Links & References
https://igniteoutsourcing.com/automotive/artificial-intelligence-in-automotive-industry/#:~:text=The%20most%20valuable%20aspect%20of,that%20learn%20as%20they%20go
https://www.adv-polymer.com/blog/artificial-intelligence-in-shipping#:~:text=Maritime%20Artificial%20Intelligence%20Industry%20Insights,-Generally%20speaking%2C%20insights&text=Logistics%20is%20beginning%20to%20become,eliminating%20mundane%20and%20routine%20tasks.
https://marine-digital.com/article_ai_and_ml
https://www.mindtitan.com/case/artificial-intelligence-in-aviation-and-travel/#:~:text=AI%20in%20the%20aviation%20industry,mistakes%2C%20and%20increase%20customer%20satisfaction.
https://www.altexsoft.com/blog/datascience/7-ways-how-airlines-use-artificial-intelligence-and-data-science-to-improve-their-operations/
https://www.investopedia.com/terms/t/transportation_sector.asp
https://www.disruptordaily.com/future-of-ai-transportation/#:~:text=%E2%80%9CThe%20future%20of%20AI%20in,transportation%20more%20accessible%20for%20all.
