Showing posts with label Driverless Cars and Trucks. Show all posts
Showing posts with label Driverless Cars and Trucks. Show all posts

Artificial Intelligence - Intelligent Transportation.

  



The use of advanced technology, artificial intelligence, and control systems to manage highways, cars, and traffic is known as intelligent transportation.

Traditional American highway engineering disciplines such as driver routing, junction management, traffic distribution, and system-wide command and control inspired the notion.

Because it attempts to embed monitoring equipment in pavements, signaling systems, and individual automobiles in order to decrease congestion and increase safety, intelligent transportation has significant privacy and security considerations.

Highway engineers of the 1950s and 1960s were commonly referred to as "communications engineers," since they used information in the form of signs, signals, and statistics to govern vehicle and highway interactions and traffic flow.

During these decades, computers were primarily employed to simulate crossings and calculate highway capacity.

S. Y. Wong's Traffic Simulator, which used the facilities of the Institute for Advanced Investigate (IAS) computer in Princeton, New Jersey, to study traffic engineering, is one of the early applications of computing technology in this respect.

To show road systems, traffic regulations, driver behavior, and weather conditions, Wong's mid-1950s simulator used computational tools previously established to investigate electrical networks.

Dijkstra's Algorithm, named for computer scientist Edsger Dijkstra, was a pioneering use of information technology to automatically construct and map least distance routes.

In 1959, Dijkstra created an algorithm that finds the shortest path between a starting point and a destination point on a map.

Online mapping systems still use Dijsktra's routing method, and it has significant economic utility in traffic management planning.

Other algorithms and automated devices for guidance control, traffic signaling, and ramp metering were developed by the automotive industry during the 1960s.

Many of these devices, such as traffic right-of-way signals coupled to transistorized fixed-time control boxes, synchronized signals, and traffic-actuated vehicle pressure detectors, became commonplace among the public.

Despite this, traffic control system simulation in experimental labs has remained an essential use of information technology in transportation.

Despite the engineers' best efforts, the rising popularity of vehicles and long-distance travel stressed the national highway system in the 1960s, resulting in a "crisis" in surface transportation network operations.

By the mid-1970s, engineers were considering information technology as a viable alternative to traditional signaling, road expansion, and grade separation approaches for decreasing traffic congestion and increasing safety.

Much of this work was focused on the individual driver, who was supposed to be able to utilize data to make real-time choices that would make driving more enjoyable.

With onboard instrument panels and diagnostics, computing technology promised to make navigation simpler and maximize safety while lowering journey times, particularly when combined with other technologies like as radar, the telephone, and television cameras.

Automobiling became more informed as computer chip prices dropped in the 1980s.

Electronic fuel gauges and oil level indicators, as well as digital speedometer readouts and other warnings, were added to high-end automobile models.

Most states' television broadcasts started delivering pre-trip travel information and weather briefings in the 1990s, based on data and video collected automatically by roadside sensing stations and video cameras.

These summaries were made accessible at roadside way stations, where passengers could get live text-based weather forecasts and radar pictures on public computer displays, as well as as text on pagers and mobile phones.

Few of these innovations have a significant influence on personal privacy or liberty.

However, in 1991, Congress approved the Intermodal Surface Transportation Efficiency Act (ISTEA or "Ice Tea"), which allowed $660 million for the creation of the country's Intelligent Vehicle Highway System (IVHS), which was coauthored by Secretary of Transportation Norman Mineta.

Improved safety, decreased congestion, greater mobility, energy efficiency, economic productivity, increased usage of public transit, and environmental cleanup are among the aims set out in the act.

All of these objectives would be achieved via the effective use of information technology to facilitate transportation in the aggregate as well as on a vehicle-by-vehicle basis.

Hundreds of projects were funded to provide new infrastructure and possibilities for travel and traffic management, public transit management, electronic toll payment, commercial fleet management, emergency management, and vehicle safety, among other things.

While some applications of intelligent transportation technology remained underutilized in the 1990s—for example, carpool matching—other applications became virtually standard on American highways: for example, onboard safety monitoring and precrash deployment of airbags in cars, or automated weigh stations, roadside safety inspections, and satellite Global Positioning System (GPS) tracking for tractor-trailers.

Private enterprise had joined the government in augmenting several of these services by the mid-1990s.

OnStar, a factory-installed telematics system that utilizes GPS and cell phone communications to give route guidance, summon emergency and roadside assistance services, track stolen cars, remotely diagnose mechanical faults, and access locked doors, is included in all General Motors vehicles.

Automobile manufacturers also started experimenting with infrared sensors coupled to expert systems for autonomous collision avoidance, as well as developing technology that allows automobiles to be "platooned" into huge groups of closely spaced vehicles to optimize highway lane capacity.

The computerized toll and traffic management system, which was launched in the 1990s, was perhaps the most widely used implementation of intelligent transportation technology (ETTM).

By placing a radio transponder on their cars, ETTM enabled drivers to pay their tolls on highways without having to slow down.

Florida, New York State, New Jersey, Michigan, Illinois, and California were all using ETTM systems by 1995.

Since then, ETTM has expanded to a number of additional states as well as internationally.

Because of its potential for government intrusion, intelligent transportation initiatives have sparked debate.

Hong Kong's government deployed electronic road pricing (ERP) in the mid-1980s, with radar transmitter-receivers triggered when cars went through tolled tunnels or highway checkpoints.

This system's billing bills supplied drivers with a full record of where they had gone and when they had been there.

The system was put on hold before the British turned Hong Kong over to the Chinese in 1997, due to concerns about potential human rights violations.

On the other hand, the basic purpose of political surveillance is sometimes broadened to include transportation goals.

The UK government, for example, placed street-based closed-circuit television cameras (CCTV) in a "Ring of Steel" around London's financial sector in 1993 to defend against Irish Republican Army terror bombs.

Extreme CCTV enhanced monitoring of downtown London ten years later, in 2003, to incorporate several infrared illuminators enabling the "capture" of license plate numbers on autos.

A daily usage tax was imposed on drivers entering the crowded downtown area.

Vehicles with a unique identification like the Vehicle Identification Number (VIN) or an electronic tag may be tracked using technologies like GPS and electronic payment tollbooth software, for example.

This opens the door to continuous monitoring and tracking of driving choices, as well as the prospect of permanent movement records.

Individual toll crossing locations and timings, the car's average speed, and photos of all passengers are routinely obtained by intelligent transportation surveillance.

In the early 2000s, state transportation bureaus in Florida and California utilized comparable data to send out surveys to individual drivers who used certain roads.

Several state motor vehicle agencies have also contemplated establishing "dual-use" intelligent transportation databanks to supply or sell traffic and driver-related data to law enforcement and marketers.

Artificial intelligence methods are becoming an increasingly important part of intelligent transportation planning, especially as large amounts of data from actual driving experiences are now being collected.

They're being used more and more to control vehicles, predict traffic congestion, and meter traffic, as well as reduce accident rates and fatalities.

Artificial neural networks, genetic algorithms, fuzzy logic, and expert systems are among the AI techniques already in use in various intelligent transportation applications, both singly and in combination.

These methods are being used to develop new vehicle control systems for autonomous and semiautonomous driving, automatic braking control, and real-time energy consumption and emissions monitoring.

Surtrac, for example, is a scalable, adaptive traffic control system created by Carnegie Mellon University, which uses theoretical modeling and artificial intelligence algorithms.

The amount of traffic on particular roadways and intersections may change dramatically throughout the day.

Traditional automated traffic control technology adapts to established patterns on a set timetable or depends on traffic control observations from a central location.

Intersections can communicate with one another, and automobiles may possibly exchange their user-programmed travel paths, thanks to adaptive traffic management.

Vivacity Labs in the United Kingdom employs video sensors at junctions and AI technology to monitor and anticipate traffic conditions in real time during an individual motorist's trip, as well as perform mobility assessments for enterprises and local government bodies at the city scale.

Future paths in intelligent transportation research and development may be determined by fuel costs and climate change consequences.

When oil costs are high, rules may encourage sophisticated traveler information systems that advise drivers to the best routes and departure times, as well as expected (and costly) idling and wait periods.

If traffic grows more crowded, cities may use smart city technologies like real-time traffic and parking warnings, automated incident detection and vehicle recovery, and linked surroundings to govern human-piloted vehicles, autonomous automobiles, and mass transit systems.

More cities throughout the globe are going to use dynamic cordon pricing, which entails calculating and collecting fees to enter or drive in crowded regions.

Vehicle occupancy detection monitors and vehicle categorization detectors are examples of artificial intelligence systems that enable congestion charging.

 



Jai Krishna Ponnappan


You may also want to read more about Artificial Intelligence here.



See also: 


Driverless Cars and Trucks; Trolley Problem.


Further Reading:



Alpert, Sheri. 1995. “Privacy and Intelligent Highway: Finding the Right of Way.” Santa Clara Computer and High Technology Law Journal 11: 97–118.

Blum, A. M. 1970. “A General-Purpose Digital Traffic Simulator.” Simulation 14, no. 1: 9–25.

Diebold, John. 1995. Transportation Infostructures: The Development of Intelligent Transportation Systems. Westport, CT: Greenwood Publishing Group.

Garfinkel, Simson L. 1996. “Why Driver Privacy Must Be a Part of ITS.” In Converging Infrastructures: Intelligent Transportation and the National Information Infrastructure, edited by Lewis M. Branscomb and James H. Keller, 324–40. Cambridge, MA: MIT Press.

High-Tech Highways: Intelligent Transportation Systems and Policy. 1995. Washington, DC: Congressional Budget Office.

Machin, Mirialys, Julio A. Sanguesa, Piedad Garrido, and Francisco J. Martinez. 2018. “On the Use of Artificial Intelligence Techniques in Intelligent Transportation Systems.” In IEEE Wireless Communications and Networking Conference Worshops (WCNCW), 332–37. Piscataway, NJ: IEEE.

Rodgers, Lionel M., and Leo G. Sands. 1969. Automobile Traffic Signal Control Systems. Philadelphia: Chilton Book Company.

Wong, S. Y. 1956. “Traffic Simulator with a Digital Computer.” In Proceedings of the Western Joint Computer Conference, 92–94. New York: American Institute of Electrical Engineers.




Artificial Intelligence - AI Systems That Are Autonomous Or Semiautonomous.

 



Autonomous and semiautonomous systems are characterized by their decision-making dependence on external orders.

They have something in common with conditionally autonomous and automated systems.

Semiautonomous systems depend on a human user somewhere "in the loop" for decision-making, behavior management, or contextual interventions, while autonomous systems may make decisions within a defined region of operation without human input.

Under some situations, conditionally autonomous systems may operate independently.

Automated systems differ from semiautonomous and autonomous systems (autonomy) (automation).

The actions of the earlier systems are preset sequences directly related to specific inputs, while the later systems' actions are predefined sequences directly tied to specified inputs.

When a system's actions and possibilities for action are established in advance as reactions to certain inputs, it is termed automated.

A garage door that automatically stops closing when a sensor detects an impediment in its path is an example of an automated system.

Sensors and user interaction may both be used to collect data.

An automated dishwasher or clothes washer, for example, is a user-initiated automatic system in which the human user sets the sequences of events and behaviors via a user interface, and the machine subsequently executes the commands according to established mechanical sequences.

Autonomous systems, on the other hand, are ones in which the capacity to evaluate conditions and choose actions is intrinsic to the system.

The autonomous system, like an automated system, depends on sensors, cameras, or human input to give data, but its responses are marked by more complicated decision-making based on the contextual evaluation of many simultaneous inputs such as user intent, environment, and capabilities.

When it comes to real-world instances of systems, the terms automated, semiautonomous, and autonomous are used interchangeably depending on the nature of the tasks at hand and the intricacies of decision-making.

These categories aren't usually defined clearly or exactly.

Finally, the degree to which these categories apply is determined by the size and scope of the activity in question.

While the above-mentioned basic differences between automated, semiautonomous, and autonomous systems are widely accepted, there is some dispute as to whether these system types exist in real systems.

The degrees of autonomy established by SAE (previously the Society of Automotive Engineers) for autonomous automobiles are one example of such ambiguity.

Depending on road or weather conditions, as well as situational indices like the existence of road barriers, lane markings, geo-fencing, adjacent cars, or speed, a single system may be Level 2 partly autonomous, Level 3 conditionally autonomous, or Level 4 autonomous.

The degree of autonomy may also be determined by how an automobile job is characterized.

In this sense, a system's categorization is determined as much by its technical structure as by the conditions of its operation or the characteristics of the activity focus.



EXAMPLES OF AUTONOMOUS AI SYSTEMS



E Vehicles that are self-driving.


 The contrasts between automated, semiautonomous, conditionally autonomous, and completely autonomous vehicle systems are shown using automated, semiautonomous, conditionally autonomous, and fully autonomous car systems.

Automated technology, like as cruise control, is an example.

The user specifies a vehicle speed goal, and the vehicle maintains that speed while adjusting acceleration and deceleration as needed by the terrain.

However, in the case of semiautonomous vehicles, a vehicle may be equipped with an adaptive cruise control feature (one that regulates a vehicle's speed in relation to a leading vehicle and to a user's input), as well as lane keeping assistance, automatic braking, and collision mitigation technology.

Semiautonomous cars are now available on the market.

Many possible inputs (surrounding cars, lane markings, human input, impediments, speed restrictions, etc.) may be interpreted by systems, which can then regulate longitudinal and latitudinal control to semiautonomously direct the vehicle's trajectory.

The human user is still involved in decision-making, monitoring, and interventions in this system.

Conditional autonomy refers to a system that allows a human user to "leave the loop" of control and decision-making under certain situations.

The vehicle analyzes emergent inputs and controls its behavior to accomplish the objective without human supervision or intervention after a goal is set (e.g., to continue on a route).

Internal to the activity (defined by the purpose and accessible methods), behaviors are governed and controlled without the involvement of the human user.

It's crucial to remember that every categorization is conditional on the aim and activity being operationalized.

Finally, an autonomous system has fewer constraints than conditional autonomy and is capable of controlling all tasks in a given activity.

An autonomous system, like conditional autonomy, functions inside the activity structure without the involvement of a human user.



Autonomous Robotics


For a number of reasons, autonomous systems may be found in the area of robotics.

There are a variety of reasons why autonomous robots should be used to replace or augment humans, including safety (for example, spaceflight or planetary surface exploration), undesirable circumstances (monotonous tasks such as domestic chores and strenuous labor such as heavy lifting), and situations where human action is limited or impossible (search and rescue in confined conditions).

Robotics applications, like automobile applications, may be deemed autonomous within the confines of a carefully defined domain or activity area, such as a factory assembly line or a residence.

The degree of autonomy, like autonomous cars, is dependent on the specific area and, in many situations, excludes maintenance and repair.

Unlike automated systems, however, an autonomous robot inside such a defined activity structure would behave to achieve a set objective by sensing its surroundings, analyzing contextual inputs, and regulating behavior appropriately without the need for human interaction.

Autonomous robots are now used in a wide range of applications, including domestic uses such as autonomous lawn care robots and interplanetary exploration applications such as the Mars rovers MER-A and MER-B.




Semiautonomous Weapons


 is an acronym for "Semiautonomous Weapons." As part of contemporary military capabilities, autonomous and semiautonomous weapon systems are now being developed.

The definition of, and difference between, autonomous and semiautonomous changes significantly depending on the operationalization of the terminology, the context, and the sphere of activity, much as it does in the preceding automobile and robotics instances.

Consider a landmine as an example of an automated weapon that is not self-contained.

It reacts with fatal force when a sensor is activated, and there is no decision-making capabilities or human interaction.

A semiautonomous system, on the other hand, processes inputs and acts on them for a collection of tasks that form weaponry activity in collaboration with a human user.

The weapons system and the human operator must work together to complete a single task.

To put it another way, the human user is "in the know." Identifying a target, aiming, and shooting are examples of these activities.

Navigation toward a target, placement, and reloading are all possible.

These duties are shared between the system and the human user in a semiautonomous weapon system.

An autonomous system, on the other hand, would be accountable for all of these duties without the need for human monitoring, decision-making, or intervention after the objective was determined and the parameters provided.

There are presently no completely autonomous weapons systems that meet these requirements.

These meanings, as previously stated, are technologically, socially, legally, and linguistically dependent.

The distinction between semiautonomous and autonomous systems has ethical, moral, and political implications, particularly in the case of weapons systems.

This is particularly relevant for assessing accountability, since causal agency and decision-making may be distributed across developers and consumers.

As in the case of machine learning algorithms, the sources of agency and decision-making may also be ambiguous.



USER-INTERFACE CONSIDERATIONS.

 

The various obstacles in building optimum user interfaces for semiautonomous and autonomous systems are mirrored in the ambiguity of their definitions.

For example, in the case of automobiles, ensuring that the user and the system (as designed by the system's designers) have a consistent model of the capabilities being automated (as well as the intended distribution and degree of control) is crucial for the safe transfer of control responsibility.

In the sense that once an activity area is specified, control and responsibility are binary, autonomous systems pose similar user-interface issues (either the system or the human user is responsible).

The problem is reduced to defining the activity and relinquishing control in this case.

Because the description of an activity domain has no required relationship to the composition, structure, and interaction of constituent activities, semiautonomous systems create more difficult issues for the design of user interfaces.

Particular tasks (such as a car maintaining lateral position in a lane) may be decided by an engineer's use of specific technical equipment (and the restrictions that come with it) and therefore have no link to the user's mental representation of that work.

An obstacle detection task, in which a semiautonomous system moves about an environment by avoiding impediments, is an example.

The machine's obstacle detection technologies (camera, radar, optical sensors, touch sensors, thermal sensors, mapping, and so on) define what is and isn't an impediment, and such restrictions may be opaque to the user.

As a consequence of the ambiguity, the system must communicate with a human user when assistance is required, and the system (and its designers) must recognize and anticipate any conflict between system and user models.

Other considerations for designing semiautonomous and autonomous systems (specifically in relation to the ethical and legal dimensions complicated by the distribution of agency among developers and users) include identification and authorization methods and protocols, in addition to the issues raised above.

The difficulty of identifying and approving users for autonomous technology activation is crucial since once activated, systems no longer need continuous monitoring, intermittent decision-making, or interaction.


~ Jai Krishna Ponnappan

You may also want to read more about Artificial Intelligence here.



See also: 

Autonomy and Complacency; Driverless Cars and Trucks; Lethal Autonomous Weapons Systems.


Further Reading

Antsaklis, Panos J., Kevin M. Passino, and Shyh Jong Wang. 1991. “An Introduction to Autonomous Control Systems.” IEEE Control Systems 11, no. 4 (June): 5–13.

Bekey, George A. 2005. Autonomous Robots: From Biological Inspiration to Implementation and Control. Cambridge, MA: MIT Press.

Norman, Donald A., Andrew Ortony, and Daniel M. Russell. 2003. “Affect and Machine Design: Lessons for the Development of Autonomous Machines.” IBM Systems Journal 42, no. 1: 38–44.

Roff, Heather. 2015. “Autonomous or ‘Semi’ Autonomous Weapons? A Distinction without a Difference?” Huffington Post, January 16, 2015. https://www.huffpost.com/entry/autonomous-or-semi-autono_b_6487268.

SAE International. 2014. “Taxonomy and Definitions for Terms Related to On-Road 
Motor Vehicle Automated Driving Systems.” J3016. SAE International Standard. 
https://www.sae.org/standards/content/j3016_201401/.




What Is Artificial General Intelligence?

Artificial General Intelligence (AGI) is defined as the software representation of generalized human cognitive capacities that enables the ...