Unmanned System Implementation Strategy - 8.4 Research Blog UNSY501

Autonomous Container Ships

Birkeland, courtesy Ong 2017

In 2020 Yara International plans to have a fully automated cargo ship in operation, called the Birkeland.  It is designed to be a zero emission, all electric 100-container ship covering a 37 mile route delivering goods.  This route will replace the current route which takes 40,000 truck routes currently (Ong, 20217).  This is a large volume of truck traffic to remove from the roads. The Birkeland will use an integrated system of GPS, radar, cameras and sensors to navigate in and out of port, along its route and for collision avoidance (Ong, 2017).

With autonomous container ships there isn’t any opportunity to spy on individuals, there is no glaring privacy issue.  There are cameras mounted to the container to help with docking and collision avoidance. This means there may be accidental recordings of people that could lead to a privacy issue.  The ships are recording people intentionally, only people in the background of a recording.  Yara doesn’t need to permanently keep any of this footage. It would make sense they need to store the footage for a short amount of time, in case of an incident they would need the footage for any following investigations.  My opinion is after two weeks the footage is reasonably no longer needed and should be deleted.  Access to the footage should be on an as needed basis, not an open source to the public. 

In the autonomous car arena there is an ethical issue on what the car should do in the event of an unavoidable accident.  Should the car save its occupants at the expense of those outside the car? Should it always save the most amount of people, possibly killing the occupants?  Would anyone buy a car that could technically be programmed to sacrifice its owner? Accidents do happen with container ships, there are cases where they have collided with other ships, docks or bridges. The automotive industry is still looking for an answer to its ethical dilemma.  One approach it is taking is to ask people to answer who should survive in unavoidable accidents, the answers are aggregated to build a set of ethics for the cars to follow.  The maritime industry can take the same approach.  Ask people to analyze an unavoidable accident and choose the best course of action.  Once enough people have been surveyed create a set of rules based on the average morals.  This is a large undertaking and may not be a perfect solution, but it’s an excellent start.

Safety is an issue for maritime operations. Docking, navigation and collision avoidance are all concerns for an automated vessel.  Rolls Royce is working on an autonomous vessel it hopes to launch between 2020-2025 (Levander, 2017).  They are developing a situational-awareness system “that integrate imagery from high-definition visible-light and infrared cameras with lidar and radar measurements, providing a detailed picture of the ship’s immediate environment” (Levander, 2017).  This information can be fed to the onboard navigation system for all docking and navigation procedures.  They will also can feed this information to remote skipper who will pilot the ship in if needed.  Manned vessels currently use the same information and systems to help.  Yara is developing their ships along the same lines, using sensors and GPS so the ship knows where it is and what is immediately around it.

In the event of loss of control or command signal there are a few options.  Both Rolls Royce and Yara will be feeding all of the information from the ship back to a remote command center, where experienced operators can take control of the ship if needed.  All safety and navigation systems on board should monitor themselves to ensure they are operational, if a system detects an issue the operators can take control bring the ship in safely (Levander, 2017).  These ships should be designed so that a pilot can board them and control the ship directly.  In the event of signal loss with the command center the ships can be programed to navigate to a predetermined point and circle until a pilot can board the craft and dock it.  Many ports currently require a local pilot to board a ship and navigate it in so this operation is currently done already.

There are many issues that still need to be worked out with autonomous container ships. Yara and Rolls Royce are both on the fore front of this development.  Both are planning to go fully autonomous in stages. Essentially manually pilot the ship at first, then let a computer navigate and steer in the open waters with crew on board, then the computer can dock and eventually over a period of years crew will not be on board (Ong, 2017).  With a slow introduction and discrete steps, issues can be identified as they occur and addressed. 

Autonomous Concept, courtesy Levander 2017

References:

Levander, O. (2017, January 28). Forget Autonomous Cars-Autonomous Ships Are Almost Here. Retrieved September 29, 2017, from https://spectrum.ieee.org/transportation/marine/forget-autonomous-cars-autonomous-ships-are-almost-here

Ong, T. (2017, July 24). The World's First Crewless Cargo Ship Will Launch Next Year. Retrieved September 29, 2017, from https://www.theverge.com/2017/7/24/16018652/first-autonomous-ship-launch-2018

Unmanned Space Exploration - 5.4 Research Blog UNSY501

In a CNN article Don Lincoln argues that space exploration should be conducted by machines first, and only conditionally by humans in the future.  He states that people are fragile “They need food, water, and air. They can exist in only a narrow range of temperatures and find inhospitable both vacuums and a radioactive environment” (Lincoln, 2017).  Since humans have a limited range they can live within, his stance is it makes sense to send machines into space for all preliminary investigations and studies of space.  I believe he has some valid points in his article.

Hubble, courtesy NASA

Since humans are fragile it’s costly to engineer spacecrafts for human habitation.  Mars Curiosity Rover costs around $2.5 billion dollars and collected large amounts of data for NASA (Lincoln, 2017). Cassini’s mission to Saturn cost around $3.2 billion and the Hubble costs around $14 billion and key in determining that the expansion of the universe is accelerating (Lincoln, 2017).  By contrast the estimated costs of a manned mission by 2030 is around $1 trillion, that dollar amount would hamstring the rest of the budget for space exploration (Lincoln, 2017). 

“The fragility of humans, our aversion for risking human life, and the all-too-human need for consumables (food, water and oxygen) require vast amounts of money to pay for the extra engineering and multiple redundant systems we demand to reduce risk to astronauts, as well as for the vastly larger support crews needed to baby-sit every aspect of daily life during a manned space mission” (Colwell & Britt, n.d.). From a budgetary standpoint I agree with Lincoln, it is significantly more cost effective to send machines to space than it is humans.  “Manned programs can cost tens or hundreds of times more than the robotic missions” (Lincoln, 2017). 

One of the ultimate goals of space exploration is to make humanity a multiplanetary species. Deciding where to live is a big question still.  Terraforming Mars, converting it to a world with and atmosphere suitable for us, could take centuries or possibly even a millennium (Warmflash, 2017).  “There is no place in our solar system where pioneers can simply drop seeds in the soil and wait for food to pop out of the ground. For that, we need to look at distant stars” (Lincoln, 2017). 

Any potential planet believed to be a suitable habitat for humans, without the need for terraforming, would need to be explored by machines first.  A trip to and from Proxima Centauri, our nearest stellar neighbor, would take around eight years (Lincoln, 2017).  Even travelling to Jupiter would take around seven years one way (The 12 Greatest, 2016).  The amount of space radiation a human would be subject to is known to cause cancer, eye issues and possibly Alzheimer’s (The 12 Greatest, 2016).  Machines do not require protections, nor do they suffer from health issues associated with space travel. 

Lincoln argues that machines should be used to explore space, for now. With the expense of sending humans to space, and the potential health effects it only makes sense to let machines do all the initial work.  Once machines have explored into the depths of space and figured out what planet would be suitable then it would make sense to start sending humans into space.  “With a welcoming destination beckoning to them, a team of intrepid men and women will leave the solar system and strike out for a new home. And, at that moment, homo interstellaris will come of age” (Lincoln, 2017). 

References:

Colwell, J., & Britt, D. (n.d.). Are robots or astronauts the future of space exploration? Retrieved September 10, 2017, from https://www.ucf.edu/pegasus/opinion/

Lincoln, D. (2017, April 20). Machines, Not People, Should be Exploring Space for Now. Retrieved September 10, 2017, from http://edition.cnn.com/2017/04/20/opinions/machines-not-people-should-be-exploring-space-opinion-lincoln/index.html

The 12 Greatest Challenges for Space Exploration. (2016, February 16). Retrieved September 10, 2017, from https://www.wired.com/2016/02/space-is-cold-vast-and-deadly-humans-will-explore-it-anyway/

Warmflash, D. (2017, September 08). Quest to colonize space demands boost from biotechnology, synthetic biology. Retrieved September 10, 2017, from https://geneticliteracyproject.org/2017/09/08/quest-colonize-space-demands-boost-biotechnology-synthetic-biology/

Future Drone Regulations - 4.4 Research Blog UNSY501

Drone use in the U.S. is becoming increasingly popular as the technology becomes cheaper and simpler to use for the novice.  The FAA expects hobbyist drones to increase from 1.1 million drones in 2016 to 3.5 million by 2021 (Shepardson, 2017).  Commercial drones are expected to increase from 42,000 in 2016 to 442,000 by 2021 (Shepardson, 2017).  The number of pilots for drones is expected to grow from 20,000 in 2016 to 200,000-400,000 by 2021 (Shepardson, 2017).  In 2016 the White House commented that by 2025 unmanned aircraft will lead to $82 billion in economic growth and support an additional 100,000 jobs (Shepardson, 2017).

With that kind of growth in the unmanned aircraft industry I believe the FAA will have to step in and develop rules governing their uses, which they have already.  The regulations will need to continue to grow and evolve with the industry. Currently part 107 states pilots need to be certified, only fly during the day, keep it in line of sight, don’t overfly people and fly in class G airspace only (Getting Started, 2017).

All of these regulations make sense and are designed to protect the operators, bystanders and aircraft in the area.  I don’t believe allow for the full growth of the industry.  If operators could operate beyond line of site it would improve the productivity of certain occupations that are using drones, such as surveying or land management. Both Amazon and Google are considering drone use for delivery of goods, which would require flight beyond line of sight and flying into neighborhoods around people (Shepardson, 2017). In a 2017 article The Economist discusses how regulation is going to drive the technology used in drones.

Drone operators and companies can apply for a waiver for part 107, provided they can show how they can still be operated safely under the waiver they are asking for (The Future of Drones, 2017).  If a drone needs to be flown after sunset it requires a light on the drone that is visible from three miles away and the operator must be trained to fly at night (The Future of Drones, 2017).  Brendan Schulman, head of policy at DJI, believes that waivers such as night flights will be what leads to new rules by the FAA (The Future of Drones, 2017).   I think it’s a step in the right direction if the FAA looks at what waivers are being applied for and shape the polices to allow for the waivers as part of the regulations.

The FAA is looking to drone operators on how to allow them to mitigate the risk of flying over people (The Future of Drones, 2017).  DJI is looking at possibly adding parachutes to drones, some form of cushioning or making them so light that if they do fall onto someone there would be little injury caused (The Future of Drones, 2017). 

DJI currently supports “geo-fencing” using AirMap.  This uses a database and the GPS in the DJI to control where drones are and are not allowed to fly, such as flying to close to an airport, the drone will not allow the operator to enter an area restricted drone flight (The Future of Drones, 2017).    If the FAA can combine current airspace restrictions and NOTAMs into something such as AirMap I feel this would be an excellent barrier to keep drones from flying where they should not. “Geo-fencing” would also allow for beyond line of sight flight, an operator can't get disoriented and fly into controlled airspace. 

The Economist points out that the regulations are what is going to drive the future of drones.  Regulations allowing for night flights will drive an anti-collision light on drones.  “Geo-fencing” will mean drones will require GPS. If some form of ATC is developed for drones they will need a version of a transponder, and a way to detect other drones in their vicinity to avoid collisions.  I personally and excited about the future of drone use and I hope to see the market grow, as well as the FAA stay current on what rules and regulations are required. 

References:

The Future of Drones Depends on Regulation, not just Technology. (2017, June 10). Retrieved September 02, 2017, from https://www.economist.com/news/technology-quarterly/21723000-engineers-and-regulators-will-have-work-together-ensure-safety-drones-take

Getting Started. (2017, July 31). Retrieved September 02, 2017, from https://www.faa.gov/uas/getting_started/

Shepardson, D. (2017, March 22). U.S. commercial drone use to expand tenfold by 2021: government agency. Retrieved September 02, 2017, from http://www.reuters.com/article/us-usa-drones/u-s-commercial-drone-use-to-expand-tenfold-by-2021-government-agency-idUSKBN16S2NM

Ocean One: a Humanoid Submarine Vessel - 3.4 Research Blog UNSY501

While researching unmanned maritime systems I discovered Stanford’s Ocean One.  A human like remote operated vehicle (ROV), a highly capable impressive machine.  What’s most interesting is that the majority of ROVs are box in shape, there is a square platform with a powertrain system and a few manipulators. This one is shaped like a mermaid with manipulators for arms. 

Traditionally electronics in ROVs are in water tight containers that keep the air in and the water out to protect the electronics. Ocean One imbeds the electronics in oil, which doesn’t compress, and gives this ROV a maximum depth of 2000 meters (Ackerman, 2016).  Ocean One has stereoscopic vision just like a person, which helps the operator view and manipulate the ROV intuitively (Feltlinger, 2016). The arms are spaced from the eyes around the same distance as our arms, it has human like proportions (Feltlinger, 2016). 

There are sensors throughout Ocean One that gauge currents and turbulence in the water, allowing for automated position control.  This means that when the operator is manipulating an object Ocean One automatically compensates and fires its thrusters to keep the itself in place, with additional thrusters in the arms to keep the hands precisely where the operator wants them (Carey, 2016).  These sensors are also used for collision avoidance.  If a collision is imminent Ocean One can use its arms to absorb the impact, just as a person would (Carey, 2016). 

The biggest advantage of Ocean One is the arms.  It enables the operator to reach out and grab objects intuitively.  Both wrists are fitted with force sensors that relay that information back to the operator, enabling them to feel how hard they are gripping objects (Carey, 2016).  There is an automated process to keep the same pressure on an object, to keep a grip on the object but not crush it (Carey, 2016).  Long term Standford wants to place sensors in each finger in order to provide more accurate feedback for the operator (Carey, 2016). 

Ocean One was originally designed to monitor coral reefs, which is why there is force feedback in the arms.  It was used in 2016 to explore a shipwreck off the coast of France. It explored La Lune, which King Louis XIVs flagship, and sank in 1664 (Feltlinger, 2016). It was able to pick up a delicate vase the size of a grapefruit and then place it in a recover basket (Feltlinger, 2016). This feat was followed by Ocean one giving a high-five to the archaeologists that dove to the wreckage with it (Feltlinger, 2016).

The press release for the archaeological survey stated “He hovered precisely over the vase, reached out, felt its contours and weight, and stuck a finger inside to get a good grip. He swam over to a recovery basket, gently laid down the vase and shut the lid “(Feltlinger, 2016).

Ocean One represents a future of unmanned vehicles that I am excited about.  The more intuitive we can make these machines the more capable they will become.  Giving force feedback to the operator and allowing them to maneuver the vehicle just like they would themselves is the right direction for unmanned systems. 

References:

Ackerman, E. (2016, April 28). Stanford's Humanoid Diving Robot Takes on Undersea Archaeology and Coral Reefs. Retrieved from http://spectrum.ieee.org/automaton/robotics/ humanoids/stanford-ocean-one-humanoid-diving-robot

Carey, B. (2016, April 27). Stanford's humanoid robotic diver recovers treasures from King Louis XIV's wrecked flagship. Retrieved from http://news.stanford.edu/2016/04/27/robotic-diver-recovers-treasures/

Feltlinger, S. (2016, August 15). Stanford Creates "Robotic Mermaid" To Help With Deep Sea Exploration. Retrieved from https://www.dogonews.com/2016/8/15/stanford-creates-robotic-mermaid-to-help-with-deep-sea-exploration

Autonomous On Highway Trucks - 2.4 Research Blog 1 UNSY501

Peloton is a company that is designing automated on highway trucks, specifically with a focus on platooning trucks. Platooning is when tracks drive in close proximity at a constant speed which reduces fuel consumption and emission (Peloton Raises, 2017).  Peloton wants to address three major issues for on highway trucks; fuel consumption, safety and operational efficiency (Peloton Raises, 2017).  

Fuel represents around 41% of the total operating costs for on highway trucks (Truck Platooning, n.d.).  Peloton and Lockheed did extensive testing along a 40 mile stretch on Interstate 80 under various conditions with tightly controlled variables such as all identical trucks and identical tire pressures (Truck Platooning, n.d.). ” The trucks carried specially manufactured fuel tanks, fed by hand pump from standard ones, that were weighed before and after each test segment and a full day’s run, using the same portable scales in all cases” (Truck Platooning, n.d.).  It was determined the lead truck saved 4.5% on fuel and the rear saved 10% when platooned (Truck Platooning, n.d.).

Peloton’s driver assist systems uses direct vehicle-to-vehicle, V2V, communication between trucks. Once paired two trucks will operate together on the highway. The trucks will match the speed set by the first truck in the pair, then all braking and accelerating is done by on board computers in unison with each other (Peloton Raises, 2017).  The computer system can be integrated into any truck, regardless of manufacturer.  This allows any trucks on the highway to platoon, regardless of manufacturer and owner of the truck (Peloton Raises, 2017).

It takes an average human driver around one second to react and apply brakes on the road, a platooned system will do this within a hundredth of a second (V2V and the Cloud, 2017). This increases the safety and allowing the trucks to travel closer together without additional risk.

The key factor behind improving safety is the V2V system.  Using radar based technology the front truck would apply the brake, moments later the truck would slow down, moments later the radar on the rear truck would detect the decrease in speed and then react (V2V and the Cloud, 2017). This reactionary process can take a second or two, which may be too long to avoid an accident.

Using V2V the lead truck is able to communicate with the rear truck when the brakes are being applied, before the lead truck even slows down (V2V and the Cloud, 2017). It will also communicate how hard the brakes are being applied, allowing the rear truck to match the braking force.

Peloton also uses cellular networks and Wi-Fi communications to feed information back to the “Cloud”, it’s main processing hub that operates as the overall control of the system (V2V and the Cloud, 2017).  If two trucks are deemed to be in a location with severe weather, platooning will be denied (V2V and the Cloud, 2017).  The trucks must be on roads designated as highways and the traffic in the area can’t be too heavy (V2V and the Cloud, 2017).  These conditions ensure that platooning only occurs under safe conditions.

One improvement Peloton is developing would be to understand the unique braking capabilities of each truck in the platoon. The truck with the stronger brakes will have to be in the rear in order to pair to improve safety (V2V and the Cloud, 2017). 

Currently Peloton’s system only control acceleration and braking, it’s an intelligent cruise control, there are no regulatory issues with this.  Peloton is looking ahead and wants to automate the rear truck more and more over time.  The front truck will take longer to automate since a driver needs to process and drive in difficult situations (V2V and the Cloud, 2017).    

References:

Peloton Raises $60 million to Improve Truck Platoon Safety and Efficiency Through Automation (April 13, 2017). Retrieved from https://venturebeat.com/2017/04/13/peloton-raises-60-million-to-improve-truck-platoon-safety-and-efficiency-through-automation/

V2V and the Cloud – Essential for Platooning (August 9, 2017). Retrieved from https://www.automotiveworld.com/analysis/v2v-cloud-essential-platooning/

Truck Platooning Trails Take to the Highways (n.d.). Retrieved from http://www.itsinternational.com/sections/nafta/features/truck-platooning-trials-take-to-the-highways/