Machine
Learning (ML) and Artificial Intelligence (AI): Drones and Self-driving Cars– Part Five
by
Dr. RGS Asthana
Senior Member IEEE
Figure 1: Amazon delivery drone [11]
Summary
This article describes
ML and AI based technology mainly related to Drones and SDCs. The apps in
Industries such as agriculture and construction, insurance and inspection,
smart cities & IoTs are described in some detail. It also briefly mentions
topics related to drones and SDCs, e.g., drones offered as a service, disaster
management, monitoring use of unauthorized drones in restricted areas; and how
to avoid as well as manage hacking & hijacking, Human error, environmental (weather, birds, collision) issues and
technical errors. The areas, such as, mob control using drones and smooth and
safe communications as well as ethical issues involved in operation of SDCs are
also mentioned.
The future of ML and AI based drone and SDC is also
discussed in ‘way forward’.
Keywords
Machine Learning (ML) Tools, Artificial Intelligence (AI), Neural Networks, Internet of Things (IoT),
Prelude
It is assumed
that one has seen [1-8] before reading this article. Multi-disciplinary
tech has made way in our daily lives now, e.g., unmanned aerial vehicles (UAV
or drones or smart flying robots) or self-driving-cars (SDCs) are, in fact, product
of a combo of robotics, artificial intelligence (AI) and many more
technologies. A huge business potential
is emerging with application of Multi-disciplinary technology in diverse areas
- such as logistics, supply chain management, transportation, search and rescue,
military, and scientific studies. Estimates suggest that the global market for commercial
drone technology applications alone, which currently stands at
about USD 2 billion, will increase to about USD 120 billion by 2020 as a result
of regulatory progress for UAV operations, effective legislation, lower costs
as well as new, innovative business models [9]. Though not directly related to
drones or even SDCs the app described below is of immense value for enhancing
crop yield and falls in category which describes apps which really are useful
to the humans. A smartphone app from
Climate Corporation – Silicon Valley based developer called Climate
Basic –
divides the entire continental U.S. into plots that are 10 meters square. The
app then gets local temperature and erosion records, precipitation expected,
soil quality, and other agricultural data to determine how to maximize yields
for each plot. This is an example of how AI is creating new methods to business
models, operations, and the deployment of people that are likely to
fundamentally change the way business operates.
Amazon recently delivered items [11] to a
customer near Cambridge, U.K. (in a drone test-flight zone) shortly after it
had received the online order. It could not be done in US as US government has
not given necessary permissions at that time. Other drone applications include
aerial photography and filming, 3D imagery, surveying/mapping, crop control,
inspections of offshore systems, pipelines, & wind towers, buildings and
verifying insurance claims.
The prototypes of
driverless cars or SDCs built by companies, such as, Google [12] (see figure 2) and Tesla
incorporate a lot of AI to free drivers from routine tasks like cruise control,
keeping in lane, and braking when the car gets too close to the vehicle in
front. GM acquired Cruise and partnered with Lyft [31].
Ford acquired Chariot and Israeli ML and computer vision startup
SAIPS. Drive.ai [10] - a Silicon Valley startup - is creating AI
software for autonomous vehicles using deep learning, which we believe is the
key to the future of transportation. Google launched a car
recently which is programmed to handle self-driving uncertainties. It uses deep
learning — a type of ML techniques to power everything in
the car from the sensors and cameras, to the vehicle’s decision-making, to the
way the car communicates with people and things around it. Apple is also
working and may launch its own SDC soon.
Figure
2: Google's fully driverless car [12]
But what happens when a SDC is involved in an accident, or when a drone
violates privacy rights? This is where ethicists, insurers, lawyers,
policymakers, transport specialists and business planners play a role. There is need to sensitize people to the
regulatory and ethical-moral issues associated with these new technologies,
e.g., when an accident is about to occur whether a SDC will have priority to
save its passenger or the pedestrian?
Drone
and SDCs: Applications
When a tornado or
hurricane comes, not only it hits infrastructure severely but also damages it. Rescue cannot be undertaken till situation
becomes safe for human workers. In such a scenario, a group of Drones could not
only be sent to the site to carry out survey
of the damage and drones can be programmed to send photos to a central point but
can also make available temporary shelter on the spot
using AI based technologies such as Additive Building Manufacturing assisted by
3D printing. All this can be done while it’s too dangerous for the human teams
to provide even an iota of help. If people need to be picked up SDCs may also
play a role if roads are serviceable else Helicopter service may be used.
Agriculture
and construction
Agriculture
[14] and construction use Unmanned Aerial Vehicles (UAVs) as these industries may
involve large or unapproachable areas of land.
Drones are extensively used in agricultural. Drones
can suggest adequate measures to formers on drought handling and also pesticide
use. Smart drones not only monitor herds
of livestock grazing in the fields but can also act as part of a coordinated
team and offer services to not only keep track of the livestock but also find
lost animals in the herds. Drones
equipped with thermal imaging can cover vast distances in a short time and
easily locate dangerous animals.
High-quality video shot
by a drone can show farmers their crops or livestock without requiring the
farmer to go to the fields. Today farmers use some time even excessive
pesticides, herbicides and fungicides on the crop irrespective of infection
caused by a particular pest or not. With
the availability of precise imagery of the crops from drone; a former may use
pest control only at locations where it is required thus not only resulting in
savings but also health crops.
One can use different
cameras to compute crop yield, or a thermal camera to detect the insufficient
irrigation issue. With proper cameras one
can even get the height of a plant from the air.
Similarly one can use drones
to measure progress in construction and client could see the progress without
visiting the site.
Construction Industry
[15] will be affected by the use of AI based technologies and will automate
purchase of construction materials and hiring engineering companies. Autonomous
TMA Truck (ATMA) is a vehicle where no driver is needed. Such vehicles may be used in various
dangerous situations. ATMA enables leader/follower truck capability that allows
the ATMA to follow a lead vehicle which is completely unmanned. Its NAV module communicates with the GPS and
collects position data called ‘eCrumbs’ [16] and passes it on to the follower
vehicle, this enables follower vehicle to move on the exact path and speed of
the leader vehicle at each point along the route. This technology is being used
by US military in the highway construction industry in difficult terrain.
ML is used to estimate
and predict building A/C energy consumption to assist in optimizing and
automating air conditioning systems. Amount of energy consumed in commercial
buildings and offices is maintained by monitoring and controlling the temperature
at a desirable level through use of artificial neural networks. This controls the cost of air conditioning.
Earthquakes cause
widespread damage and ongoing aftershocks result in losses in billions of
dollars. With help of AI and low cost remote sensing data one can detect
building collapse in post-earthquake environment. Precise
plotting of damage can help in decision making that when people can return
safely to their home. One can also know about delayed building collapses and
save lives.
Insurance and inspections
In insurance industry drones
can play a major role. Insurance companies get a lot of requests for roof
inspections due to damage from natural disasters like hail, wind or ice. Instead
of having a person climb up on the roof, which is dangerous, they can now send
a drone to capture images of the damage and decide on the claim by analyzing
the imagery. AI software can play a major role here.
Claims
management [14] can be augmented using ML and AI techniques in different stages
of the claim handling process. By leveraging AI insurers can not only
fast-track certain claims by reducing the
processing time but also
identify patterns in the data
leading to identification of fraudulent claims in the process. This in turn will reduce the handling costs
and also enhance the customer experience. AI systems, with their self-learning
abilities over time can improve with experience and locate new unseen cases and
further improve the fraud detection. Furthermore, machine learning models can automatically
assess the severity of damages and predict the repair costs from historical
data, sensors, and images.
With AI, there will be a change in the skills needed to work in the
insurance industry. Tech savvy experts must train for the future and learn to
work together with AI to show their true innovative potential.
Smart
cities and IoT
The infrastructure may
include smart communities, connected homes, intelligent transport systems,
e-health, e-government, e-education, smart grids and smart energy
solutions. Smart cities [21] will use latest
AI based technologies and may also have Electric Vehicles, SDCs, Mobile
applications, Drones, Wearable and Smart devices and so on are just some of the
key developments to watch. The data may be collected from IoT [6]
intelligent devices and then decisions taken and solutions implemented
automatically by AI or ML based algorithms, e.g., a drone may be sent not only
to monitor status of power line but also to replace fused bulbs.
Disaster
Management: Rope bridges
A drone or group of
drones may not only provide data, but use AI and ML techniques for disaster
management by physically augmenting the infrastructure around us, that will be
the most cutting-edge usage of UAV technology. One such example could be that
Drone sets up a rope bridge say across a river on a stormy weather day. It will however need special hardware,
software, battery life, and also FAA regulation (in USA) to implement. Drones may also become a part of IoT [6].
Drones
offered as a service
The Skymatics app [18] is available on
Google Playstore and clients can book UAV services and jobs instantly on their
mobile device.
Figure 3 describes
architecture of a Drone which works on voice instructions as well as can
identify faces. This Drone was built at
a cost of less than USD 200 [19]. The Azure
Face API [20] lets you upload pictures of your friends and it recognizes them. It also gives face attribute like age,
emotion, gender, pose, smile, and facial hair along with 27 landmarks for each
face in the image.
Monitoring
unauthorized Drones [24]
A weatherproof device called DroneTracker, has number of sensors including
a near-infrared cameras, acoustical sensors and video camera to provide a
120-degree arc of airspace coverage at a range of up to 100 meters. It sends data
in real-time to a paired device for viewing and to create a temporary
drone-detecting perimeter at public events or in restricted and sensitive areas.
Figure
3: Architecture of Drone which takes
voice commands and also recognizes faces
Avoid remote hacking and
hijacking [25]
Jonathan Andersson – a
Security researcher - has developed a tiny gadget called, ‘Icarus’ [26] - a
name based on Greek mythology - to enable hijacking
nearly any drone mid-flight. This device
allows attackers to lock the owner out and give them complete control over the device.
This device [26] can, in fact, affect any radio controlled device using the
popular DSMx radio platform.
DSMx is an evolution
of DSM2 2.4 GHz technology [28], destined to deliver the most robust, reliable,
and efficient level of RF link possible in the modeling world today. The hack
works against any drone that communicates over DSMx, a widely used remote
control protocol [27] for operating hobbyist drones, planes, helicopters, cars,
and boats.
Human error, environmental
(weather, birds, collision) issues and technical errors
It is expected that there
will be more than 25 billion IoT [6] connections in the next 10 years and drone
technology will be part of the connected world.
Although this is an interesting technology but it has all risk factors
of aviation.
Luckily there are
established norms for the aircrafts which has humans; the norms for Drones will
also evolve with time as it offers an unmanned scenario. Human errors as well as natural disasters have
played a major role for flying safely in the skies. We can minimize technical
as well as human errors with training but natural disasters are beyond our
control.
Monitoring Mob
control using Drone Technology
The police in the capital of the U.P. have been
using drones to take aerial photographs since 2014, but new drones ordered [29]
by the force have the capability to control crowds using pepper spray and also fire up to 20 paintballs per second
while simultaneously dispersing tear gas pellets onto people. The drones are also fitted with onboard
speakers so authorities can communicate with crowds, as well as bright
strobe lights and “eye safe” lasers to disorient and disperse a gathering.
Dubai [32]
police is now using drones with 60 minutes flight time to monitor
crowds at soccer matches.
How to manage safe
and smooth communications
Drone technology is
exciting but very challenging and complex. Time will tell to what extent is its
wider applications are realistic? The issues related to the management of cellular
connectivity for coordinated operation and control of drone [30] as well as
SDCs enabling a growing set of use cases within and beyond the operator’s
visual line of sight. In fact, connectivity is important for safe operation of
both drone and SDC.
SDCs and ethics [23]
Driverless
cars will be fully accepted someday. The
ethical questions need be answered before that day could be:
1) Who will decide who
lives or who dies during an accident scenario say, between a passenger in the
car or a pedestrian on the road?
2) Whether a
driverless vehicle could choose to ram a school bus full of kids or sacrifice
the driverless vehicle's occupants during a mishap?
3) Should police be
given control to remotely control a SDC to pull it over?
Way
forward
The wave of progress encompasses all aspects of computer
technology, software engineering and is applied to business [17]. Companies today
have to decide what type of AI role they should play. The ML and AI based technology
used in a Drone, robot, SDC or a truck is, particularly, fascinating.
AI progress, in fact, will affect all technologies described above but will
surely be used to further improve navigation of drones and SDCs.
Drive.ai was founded to develop technology with the potential to save
lives and transform industries. The Google car [12] - a SDC - is round, white and gray
car. It is without permanent driving
setup like a gas pedal or wheel. However, to comply with California state law,
there are still removable, temporary controls for the required "safety
driver" -- a real person who needs to be in the car and ready to take over
in an emergency. The goal is to eventually remove any interior controls so that
passengers can take a nap or knit while the car does all the work.
Connected and autonomous vehicles in general are viewed as the future of
surface transportation, and this technology may be one of the first ways in
which it gets commercialized.
A recent report
from the World Economic Forum (WEF) found that the digitalization of telecommunications
could unlock $2 trillion of value over the next decade – value for the
telecommunications industry and society as a whole.
AI, Big data and IoT [6]
are transforming how future wars will be fought. Drones [33] in military are used for a
variety of purposes including conducting surveillance, for attacking hostile
targets, spotting submarines and mines, and for delivering aid to places where ii
is unsafe for human convoys. This all is possible as drones are small, can be
stealthier and safer for the troops employing them. The likelihood of seeing a drone
from the ground is very remote. Drones are activated and operated from a
computer station located in friendly territory.
Major automakers like Tesla, BMW, Ford,
Audi, and even Google are in the SDC game in some way or other [34]. While many
have only started hearing about autonomous technology recently, research is going now for decades. In fact, AI
is the brain of SDCs. Each autonomous vehicle has advanced tools to gather
information, including long-range radar, LIDAR, cameras,
short/medium-range radar, and ultrasound.
Apple Inc. U.S.A. is strongly rumored to be in the race for SDCs. AI can assist a Chauffeured car in the
following tasks:
·
Directing
the car to a gas station or recharge station as applicable;
·
Finding
the fastest route;
·
Incorporating
speech recognition for advanced communication with passengers;
·
Locating
parking space and park by itself or provide assistance in parking;
·
Eye
tracking for improved Chauffeur monitoring; and
·
Provide
intelligent user interface by using NLP and/or virtual assistance technologies.
Future of autonomous cars as well as
SDCs will depend on acceptance of advanced AI algorithms by the government as
well as public to allow them on public roads permanently.
References
[1] Progress and
Perils of Artificial Intelligence (AI)
[2] Invited Chapter 6 - Evolutionary Algorithms and Neural Networks, Pages 111-136, R.G.S. Asthana
in book, Soft Computing and Intelligent Systems (Theory and Applications),
Academic Press Series in Engineering, Edited by:Naresh K. Sinha, Madan M. Gupta
and Lotfi A. Zadeh ISBN: 978-0-12-646490-0
http://www.sciencedirect.com/science/book/9780126464900
[3] Future 2030
by Dr. RGS Asthana, Senior Member IEEE
[4] Machine
Learning (ML) and Artificial Intelligence (AI) – Part 1, by Dr. RGS Asthana, Senior Member IEEE
[5] Machine Learning
(ML) and Artificial Intelligence (AI) – Part Two, by Dr. RGS Asthana, Senior
Member IEEE
[6] Internet of Things (IoT)
[7] Machine
Learning (ML) and Artificial Intelligence (AI): Cognitive Services and Robotics
– Part Three by Dr. RGS Asthana, Senior Member IEEE
[8] Machine
Learning (ML) and Artificial Intelligence (AI): Big Data and 3 D Printing
– Part four by Dr. RGS
Asthana, Senior Member, IEEE.
[9] Drones, AI and
getting undergrads ready for great disruption
[10] Drive.ai website
[11] The first Amazon PrimeAir drone delivery landed 13 minutes after the
order was placed
[12] Google's fully functional driverless car is adorable
[13] Best Agricultural Drones of 2017 – Reviews and Specs
[14] How
Artificial Intelligence is changing the Insurance Business
[15] Boosting Construction Industry with AI
[16] First autonomous construction trucks to hit the roads
[17] A Strategist’s Guide to Artificial Intelligence
[18] Skymatics - Drone Service App
[19] How to
build an autonomous, voice-controlled, face-recognizing drone for $200
[20] Microsoft
Azure: Face API
[21] Global Smart Infrastructure - Smart Cities and Artificial
Intelligence the Way Forward
http://www.prnewswire.com/news-releases/global-smart-infrastructure---smart-cities-and-artificial-intelligence-the-way-forward-274656431.html
[22] Drones: Applications of the future
[23] Section of Science &
Technology Law: AI and Robotics Committee
[24] Jay science Tech website:
DroneTracker Detects Unauthorized Drones
[26] Hacker's Icarus machine steals drones midflight
[27] There’s a
new way to take down drones, and it doesn’t involve shotguns
https://arstechnica.com/security/2016/10/drone-hijacker-gives-hackers-complete-control-of-aircraft-in-midflight/
[28] Horizon website: DSMx
[28] Horizon website: DSMx
[29] INDIAN POLICE BUY PEPPER
SPRAYING DRONES TO CONTROL ‘UNRULY MOBS’
[30] Cellular enables safer drone
deployments
[31]Lyft
[32] Dubai debuts drones for crowd control
[33]
How AI, Drones And Big Data Are Reshaping the Future Of Warfare
[34] How AI is driving the future of autonomous cars
No comments:
Post a Comment