Computer Vision and Deep Learning for Transportation 2
Transportation is an important part of daily life as it enables the movement of goods from one place to another, trade, commerce and communication to build civilization. The transportation sector has gone through many revolutions over the past hundred years. Today we are at the stage of achieving major breakthroughs in transportation through artificial intelligence (AI).
Artificial intelligence is already transforming the transportation industry, enabling cars, trains, ships and planes to autonomously automate and make traffic flow smoother. In addition to making life easier, it can provide a safer, cleaner, smarter and more efficient mode of transportation for everyone. For example, AI-led autonomous transportation could help reduce human error in many traffic accidents.
In 2017, the global transportation-related artificial intelligence technology market reached US$1.2 billion to US$1.4 billion. Furthermore, it is expected to grow to US$3.5 billion by 2023, with a CAGR of 12-14.5%.
In addition, the accessibility and affordability of high-end computing hardware (such as CPUs and GPUs) and IoT technologies (such as LTE and 5G) create possibilities for multiple applications of artificial intelligence in transportation. This series on CV and DL for industrial and large enterprise applications will cover the benefits, applications, challenges, and trade-offs of using deep learning in transportation.
This lesson is Part 2 of 5: CV and DL 102 for Industrial and Enterprise Applications.
- Computer Vision and Oil and Gas Deep Learning
- Computer Vision and Deep Learning for Transportation (This Tutorial)
- Computer Vision and Logistics Deep Learning
- Computer Vision and Deep Learning for Healthcare
- Deep learning for computer vision and education
1. Benefits
1.1 Safety and reliability
Human error (such as speeding, distracted driving and drunk driving) causes more than 90% of road accidents. Therefore, safety and reliability are arguably the most important factors for anyone working in the travel or transportation industry. Passengers need to know that they and their belongings are safe and that the vehicle or operator they are traveling with is reliable.
With the emergence of artificial intelligence technology, security levels can reach higher peaks. The ability of AI to analyze large amounts of data will enable travel and transport operators, and ultimately the public themselves, to arrange public and private transport services in significantly improved ways. Researchers believe that by 2050, artificial intelligence cars can reduce traffic deaths by 90% in some developed countries.
1.2 Efficiency
Developing countries rank lower than developed countries onLogistics Performance Index (LPI) due to insufficient infrastructure and poor customs procedures leading to inefficiencies. Artificial intelligence will undoubtedly improve the energy efficiency of transportation systems, an important aspect of mobility. In particular, artificial intelligence can helpe-logistics and apply Internet-related technologies to the supply and demand chain ( the supply and demand chain ), matching shippers with express service providers.
1.3 Pollution
Emissions from transportation vehicles are a significant contributor to increased pollution and global warming. As the world becomes increasingly environmentally conscious, emissions from the tourism and transport industries need to be significantly reduced to ensure their long-term sustainability . Artificial intelligence can help scientists discover greener ways to operate vehicles and machinery, thereby deploying new innovative solutions to combat growing pollution. McKinsey predicts that self-driving trucks will reduce operating costs by approximately 45%. Therefore, the environmental impact will also be greatly reduced.
2. Application
2.1 Road Transport
Road transportation is one of the most important areas where artificial intelligence is currently widely used. Globally, automakers and technology companies are exploring various artificial intelligence technologies and algorithms to develop smart cars for commercial and personal use. Such vehicles use a variety of sensors,such as GPS, cameras and radar, in combination with actuators (devices that convert input signals into actions), control units and software to perform intelligent actions (e.g. , automatic driving).
2.11 Truck platooning
Truck marshalling refers to connecting several heavy goods vehicles (HGV) within a minimum distance so that they can accelerate or brake. A human drives the lead heavy goods vehicle, and only in complex traffic situations or unexpected events will the driver of the following heavy goods vehicle be present. It is expected that with the intervention of artificial intelligence, the driver's responsibilities in subsequent heavy goods vehicles will gradually be reduced.
2.12 Traffic Management
Artificial intelligence algorithms are currently being used extensively in vehicles to improve traffic flow in the economy. For example, Short-distance ride-sharing platform Uber uses AI technology in every aspect of its service, from matching customers and drivers to route optimization. Artificial intelligence algorithms are also being used in road traffic management to analyze traffic patterns and volumes and provide drivers with information on the fastest routes. This relieves their traffic congestion and reduces their commute time. This algorithm also keeps traffic flowing with real-time rotating traffic lights and signals to meet surface traffic flow needs.
For example, Surtrac, a subsidiary of Carnegie Mellon University's Rapid Flow Technologies, coordinated a network of nine traffic lights on three major roads in Pittsburgh to install traffic control solutions. Their solution reduced travel times by 25% and waiting times by an average of 40%. Vehicle emissions have also been reduced by 20%.
2.2 Aviation
The aviation industry is no stranger to the adoption of artificial intelligence technology. However, with recent advances in artificial intelligence, the industry will have a significant impact on the way its business is conducted. Managing growing air traffic is an important area for advances in automation and computing power. The International Air Transport Association (IATA) report points out that using enhanced computing power to develop unmanned aircraft systems systems (UAS) and drone traffic management will improve existing traffic management systems, separation standards andairspace planning design< /span> Create new opportunities.
The Single European Sky ATM Research (SESAR) joint project supports several artificial intelligence research projects and air traffic management ATM on issues including:
- Predictability of traffic at different flight phases
- Passenger flow improvements at airports
- greater automation
One potential application of machine learning in the airline industry is to translate historical and real-time insights into customer behavior into real-time strategies (e.g., adjusting website content delivered to customers). This also includes social media sentiment analysis, which predicts customer needs based on their social media behavior. For example, the BigData4ATM project looks at how AI can analyze disparate passenger-centric geolocation data to identify patterns in passenger behavior, door-to-door travel times and travel mode choices (Figure 5). Through smart devices and related services, researchers can access large-scale, detailed longitudinal dynamic data to test hypotheses about passenger behavior.
Another area where artificial intelligence can help is in ground handling. This includes high potential usage situations such as safety inspections, aircraft movement operations (pushback and towing), aircraft turnaround operations (refueling, catering, loading and unloading, de-icing and anti-icing) ) as well as ground transportation (passengers, baggage, cargo and mail) on ramps. It can also promote airport security by digesting large amounts of historical and real-time data.
In addition, aircraft manufacturers are using artificial intelligence to solve problems faced by pilots in the cockpit and predictive maintenance of aircraft. For example, British startup Aerogility is working with easyJet to automate its fleet's routine maintenance program, including predictive overhauls, engine shop visits and landing gear overhauls. The Airbus manufacturer uses a similar tool, Skywise, to provide predictive maintenance and data analytics.
Several companies are also beginning to use artificial intelligence technology in drones to deliver various sizes and types of goods. For example, California-based startup Nautilus is developing a 90-ton cargo drone.
2.3 Railway Transport
Railways were one of the most innovative sectors of the economy, allowing passengers to travel long distances for the first time. They were also an important factor in the Industrial Revolution. Artificial intelligence can help improve manufacturing, operations and maintenance for rail operators and infrastructure managers. It will improve management, reduce costs and increase competitiveness against direct competitors or other modes of transportation.
2.3.1 Intelligent Train Automation
One of the most important contributions of artificial intelligence in rail transportation is the automation of train operation (ATO) which shifts the responsibility for managing train operations from the driver to train control systems with varying degrees of autonomy. This typically involves transferring the train driver's senses and intelligence to the autonomous driving module, preparing to react to possible hazards. Additionally, design modules should understand and integrate passenger behavior on railway platforms, allowing train doors to close automatically without danger.
Shift2Rail is a joint EU research and development project that is developing and validating standard ATOs for all railway segments (mainline/high speed, urban/suburban, regional, and freight lines). In addition, they are carrying out ATO-related activities to optimize resource utilization. As the container shipping market develops, many projects are being developed to better synchronize the movement of container trains, thereby improving real-time information and data exchange.
2.3.2 Operational Intelligence
In railways, it is valuable to identify and be aware of potential faults before they occur to avoid service disruptions. Today, AI can use data provided by sensors (placed at strategic locations) to extract valuable insights and information and recommend maintenance actions. This will help operators reduce the amount of fleet reserves they need to keep in the event of a breakdown, as AI enables them to increase reliability and effectiveness. Other benefits include:
- "Instant forecasts" and predictions of infrastructure or rolling stock conditions
- Faster, less comprehensive repairs
- Reduce maintenance costs
- better customer satisfaction
To illustrate some AI applications, the French National Chemical Industry Society (SNCF) has started predictive maintenance on pantographs ,The pantograph may become brittle due to wear and tear. The company says that over time it will be able to predict that 80% of accidents will occur on the connections that power trains. SNCF said predictive maintenance has also reduced accidents involving train switches by 30%, and this technology has been applied to many train systems and subsystems.
SNCF railway predictive maintenance solutions include the following:
- The train communicates independently with Internet of Things IOT when passing by ground instruments
- Transfer data to data storage server
- Scientists and rail experts analyze data
- Once the data has been analyzed, it is retrieved for use by computerized maintenance management software
2.3.3 Asset Intelligence
Artificial intelligence in railways can also be used to assess the long-term performance of railway assets and recommend areas for improvement in their product design. Artificial intelligence can analyze data generated by railway infrastructure and train subsystems, helping equipment manufacturers build digital representations of physical entities, called digital twins. This enables IT, operations and engineering to access the overall picture and understand asset degradation, failures and customer behavior. These improvements represent a competitive advantage for equipment manufacturers and rail operators.
International engineering firm Laing O’Rourke uses artificial intelligence and asset digital twins to schedule maintenance work, enabling the company to reduce scheduling activities to 19 seconds.
2.4 Shipping, Navigation and Ports (Shipping, Navigation And Ports)
Ship traffic has grown significantly over the past few decades, increasing stacks of maritime safety. Additionally, as container traffic in ports increases, there is a need to accommodate port terminals and better connections. The growing size of ships increases the pressure they place on ports and their cities.
In addition, growing environmental concerns have prompted the industry to adapt to greener rules amid fierce international competition in the global maritime industry. Artificial intelligence can analyze information to provide insights for better decisions, improve safety and energy efficiency, and optimize logistics.
2.5 Maritime Shipping and Inland Navigation
Maritime operations need to adapt to changing conditions and make decisions based on several parameters. Data from advanced navigation systems (such as radar, electronic navigation charts, auto-pilot systems, wave radars, oil spill detectors, and other sensors) Analyze through artificial intelligence algorithms to extract insights and perform technical operations and maintenance. Automatic Identification System AIS transmits data (for example, ship ID, position, course, speed and destination).
By combining recorded ship movements with an advanced image recognition system, ships can be identified even when AIS is turned off. This can detect anomalies in offshore operations and improve sea level safety. In addition,machine learning algorithms can predict delays caused by severe weather conditions and traffic congestion and estimate future demand for oil prices.
Autonomous ships are an obvious application of artificial intelligence in this area. For example, the EU-funded Maritime Unmanned Navigation through Intelligence in Networks (MUNIN) project developed and tested the concept of autonomous merchant ships, which are mainly guided by automated decision-making systems on board but controlled by remote operators on shore. . In addition, the vessel uses independent operations during deep-sea voyages, rather than in congested waters or during port calls.
2.6 Port
As the amount of information from large ports around the world continues to increase, a new application called port call optimization has become very popular. It combines IoT, cloud computing and geographic information to optimize port operations, increase productivity and improve relationships with customers. In addition, the analyzed data can be used for forecasting and real-time planning to improve decision-making and support the port’s economic growth.
By using more advanced digital technologies, ports can become “smart”:
- Provide a seamless supply chain
- Optimize the allocation of relevant resources, services, and supervision
- Autonomous loading and unloading (which containers are unloaded first and how they are stacked)
- equipment scheduling (optimizing the use of cranes and vehicles)
- berth availability planning
Ports recognized as the most intelligent include Singapore, Rotterdam, Tianjin and Dubai.
Data collected during container port operations is stored and analyzed as the basis for future artificial intelligence-assisted tools that are expected to one day be able to manage the entire delivery cycle and further optimize terminal operations. Furthermore, these technological advances are understood as part of a broader supply chain transformation.
3. Challenge
There are several risks and challenges associated with artificial intelligence in transportation. However, these risks and challenges can have significant socioeconomy effects and must be managed.
Loss of Jobs: More than 4 million people will lose their jobs as the U.S. rapidly transitions to autonomous vehicles, according to a report from the Center for Global Policy Solutions. These jobs include drivers for delivery, heavy truck drivers, bus drivers and taxi drivers. In addition, artificial intelligence may accelerate the growth towards a server-based economy, which will accelerate the unemployment of low-skilled workers.
Investment: An important constraint on the growth of artificial intelligence in the transportation market is its high investment nature in terms of talent, hardware, and software. The demand for artificial intelligence experts is growing dramatically, and the lack of skilled artificial intelligence talents will become the biggest obstacle to the adoption of artificial intelligence in developed countries. In addition, huge investments and capital stock are required to leverage technical personnel and business practices.
Poor and Underdeveloped Infrastructure: Low-income and developing countries like India face huge challenges in adopting AI because of the low quality of their infrastructure. This includes roads, ports, maintenance and repair depots. Lack of reliable power and weak telecommunications make things even more difficult. Harnessing the power of AI may be more challenging in countries with little investment in technological research and complex infrastructure as a share of GDP.
Regulatory Requirements and Privacy Concerns: Regulatory requirements for artificial intelligence are always difficult to predict. Although studies show that electric vehicles can reduce traffic fatalities, it remains unclear who will bear ultimate responsibility if an accident, injury or death occurs. Likewise, requiring users to provide personal data to develop powerful machine learning models requires privacy laws. These laws must be balanced against the benefits of having more data in telecommunications networks.
Summarize
Here’s how AI can deliver safer, cleaner, smarter, and more efficient transportation for everyone.
- Road transportation: Road vehicles can use a variety of sensors, such as GPS, cameras and radar, and actuators (which will enter devices that convert signals into actions), control units and software to perform intelligent actions such as autonomous driving.
- Aviation: Translate historical and real-time insights into customer behavior into real-time strategies (e.g., adjust website content displayed to customers).
- Rail: Artificial intelligence can help improve manufacturing, operations and maintenance for rail operators and infrastructure managers. It will improve management, reduce costs and increase competitiveness against direct competitors or other modes of transportation.
- Marine and Shipping: Data from advanced navigation systems such as radars, electronic navigation charts, autonomous driving systems, wave radars, oil spill detectors and other sensors are analyzed through artificial intelligence algorithms to extract insights and enable technical operations and maintenance.
However, artificial intelligence also faces challenges in the transportation field.
- Unemployment: Artificial intelligence is likely to accelerate growth toward a service-based economy, which will accelerate unemployment among low-skilled workers.
- Investment: A significant constraint on the growth of artificial intelligence in the transportation market is its high investment nature in terms of talent, hardware, and software. The demand for AI experts is growing dramatically, and the lack of skilled AI talent will pose the most serious obstacle.
- Weak and underdeveloped infrastructure: Low-income and developing countries like India face huge challenges in adopting AI because of their low-quality infrastructure. This includes roads, ports, maintenance and repair depots.
- Regulatory requirements and privacy concerns: Regulatory requirements for AI are always difficult to predict. Although studies show that electric vehicles can reduce traffic fatalities, it remains unclear who will bear ultimate responsibility if an accident, injury or death occurs.