Autonomous vehicles are being utilized across numerous industries resulting in dramatic increases in business performance. To operate safely and successfully, these autonomous vehicles must navigate their surroundings without incident. As such, collision avoidance is one of the largest areas of concern when it comes to the implementation of autonomous vehicles.
Accurate and reliable detection of obstacles is paramount to collision avoidance, but can be quite challenging as obstructions tend to vary in size and geometry. Some objects, such as dust, can cause a false-positive or false-negative reaction, when in reality, there is no impending collision.
The problem of obstacle detection can be reduced to a two-state Markov model, which is a model used for randomly changing systems based on stochastic calculus. The transition between these two states, obstacle or non-obstacle, is then governed by a Bayesian inference which takes into account past probability, current measurements, and sensor characteristics. Thus, the most comprehensive way to correctly understand the real obstacles autonomous vehicles encounter is by applying sensor fusion.
By “fusing” LiDAR and Radar measurements into a complete occupancy grid, along with advanced Artificial Intelligence (AI) and Machine Learning (ML), ASI’s Obstacle Detection Sensor Fusion application optimizes obstacle detection for increased collision avoidance. Through the implementation of predictive analytics which account for prior inputs of cell occupancy probability, a labeled 3D point-cloud from the LiDAR unit, the detected objects by the radar, and the associated radar field of view, ASI’s algorithm builds a complete grid system that identifies, classifies, and circumvents obstacles.
The value in each grid cell is computed using the measurements that are classified as either obstacle or non-obstacle. Then, based on adjustments to both true-positive and true-negative rates, the AI algorithm is tuned for each sensor by using the associated confidence level. This method has been show to enhance collision avoidance.
ASI’s Obstacle Detection Sensor Fusion enables autonomous vehicles to traverse unknown terrain and dangerous obstacles to arrive at a desired endpoint safely. The advanced AI and ML algorithm creates an accurate model of the world, identifies and classifies obstacles and non-obstacles, then stores this information in a database which can be accessed by any autonomous vehicle that enters that region, thereby allowing for continuous updating and learning. Teaching a vehicle to learn is just the beginning — ASI’s research and development team is taking this technology further by enabling autonomous vehicles to share information for a collaborative improvement in vehicle behavior.
Learn more about ASI’s collision detection at work in the mining industry and current areas of research.
Elements of driverless technology are being inserted into every industry and sector, from consumer vehicles to commercial transportation to mining fleets.
The military is also actively pursuing ways in which ground vehicle operations can be made safer and more efficient with the integration of autonomous solutions. One way in which this is possible is by transforming manned convoys into unmanned operations.
A Smarter Convoy
The military relies on convoys to move people and supplies, typically along vast and sometimes treacherous terrain. Obstacles, enemy combatants, even inclement weather can pose challenges that might stand in the way of a successful convoy deployment. However, artificial intelligence and advanced robotics allow convoys to create improved, safer, and more reliable paths.
ASI’s Autonomous Convoy System (ACS) uses innovative technology that enables convoy vehicles to communicat with one another, relaying information about obstacles, terrain or other unsafe conditions. While numerous technology companies have made investments in autonomous convoy development for military applications, The United States Army Tank Automotive Research, Development and Engineering Center (TARDEC) turned to ASI to help advance current autonomous convoy technology.
Our research group has engineered breakthroughs for lateral stability control improvements. This allows leading vehicles to pass information back to followers, which the following vehicles use to improve their path following. We’ve also made advancements in sequential stability, multiple vehicle convoy, and world modeling, as convoys often encounter unmapped territories.
ACS from ASI allows convoys to chart the smartest path to a destination. Autonomous Ground Resupply (AGR) provides another solution for generating better efficiencies. With AGR, troops can go out into the field with autonomous supply vehicles following them. When supplies diminish, the vehicles can return to base and retrieve more of what the field troops need. This saves time and costs, not having to transport human personnel back and forth across possibly dangerous supply lines, allowing them to continue executing their mission.
The advantages ACS yield for military applications can also be transferred to other industries. Convoys of unmanned vehicles can be useful for farming, shipping, and industrial organizations, reducing costs, decreasing personnel, and saving time by using smarter paths. To learn more about ACS and how ASI technology platforms are automating vehicle transportation, visit us at asirobots.com/platforms.
ASI is working on things that no one else is doing today—taking existing vehicles and converting them to autonomous. We can use the same basic kit on a lot of different vehicles: mining equipment, agriculture vehicles, and everything in between.
This makes it possible to reduce driver exposure to hazardous conditions. After all, a robot can’t get a back or neck injury.
The Mobius technology has been developed over the last 17 years to optimize systems, so an operation can get lots of vehicles working together. One person can control 50 vehicles, and the Mobius technology keeps them from hitting each other and enables them to run constantly for 24 hours.
Projects that were never financially viable now become a possibility under an autonomous operation. The ASI system can also interface with vehicles that are not autonomous, which gives operators the ability to work in a mixed traffic environment.
Putting the robot in a vehicle opens the door for many industries to be more innovative and a lot safer. If you can dream it, there are people at ASI who can do it. To see the game-changing capability of ASI’s solutions in action, enjoy the video.
SALT LAKE CITY, JULY 13, 2017 – Autonomous Solutions, Inc. (ASI) has been awarded additional funding through the Department of Defense (DoD) to further apply machine learning and artificial intelligence to improve the mobility and behavior of autonomous vehicles in challenging environments.
Two other ongoing programs with the government include the development of machine learning for obstacle classification, LIDAR-camera fusion, and vehicle auto tuning using artificial intelligence. The developments will enable ASI’s autonomous ground vehicles to adapt in real time to difficult dynamic environments for automotive, agriculture, mining, construction, floor cleaning, security surveillance, and lawn mowing robots.
“In order for us to maintain our leadership in the unmanned vehicle space we must continue to push the boundaries of what is possible with machine learning and artificial intelligence,” says CEO and founder of ASI, Mel Torrie. “These programs will ensure that we continue to offer our customers and partners the most advanced safe and simple autonomous solutions.”
“The ability for vehicles to learn from past experience and continuously improve as they drive is important. These benefits multiply as the robots share their learning with other vehicles in the area in real time. The use of Machine Learning and AI will enable huge strides in efficiency improvement and maintenance reduction,” says Dr. Jeff Ferrin, Research and Development General Manager.
Mel Torrie will be sharing more about this exciting technology development at the Silicon Valley Innovation and Entrepreneurship Conference in Beijing China on 15 July 2017 and at the Silicon Valley-China AI & Investment Forum on 19 July 2017 in Santa Clara, California in an AI session alongside leading pioneers in this space from companies like Tesla, IBM, Facebook, and Samsung.
In the proving ground environment, car manufacturers engage in extensive durability and misuse testing.
They might jump a vehicle off a ledge, run it into a wall, or even run tests in which the vehicle rolls over—all to improve the safety of the vehicle for consumers.
The challenge is there is a clear limit to what you can test with a driver in the car. Though the ultimate buyer of the vehicle may misuse a car or truck in hundreds of imaginative ways, proving grounds need to keep the health and safety of their drivers in mind when they’re designing tests.
In fact, one of ASI’s customers recently had to shut down a ditch twist test the first day they started running it, after two test drivers endured neck injuries. Since then, they turned to ASI to automate their vehicle. They now run the test all the time, ultimately improving the safety of the vehicle itself.
Putting the robot in a vehicle opens the door for a proving ground to be very innovative in what kinds of tests they can write and run. To learn more about the capability of ASI’s proving ground solution, please enjoy the video.
For automakers, manufacturing vehicles that are safe is the number one priority. In order to evaluate the safety and reliability of cars and trucks, companies test their models on miles of tracks known as proving grounds.
Here, human drivers are often employed to subject automobiles to rigorous handling and varying conditions so engineers can assess the effects of different speeds, driving scenarios, and varying passenger loads.
But the human element of these durability and misuse tests comes with many disadvantages that, when removed, can optimize operations leading to faster results and direct cost savings.
The first advantage of replacing a human driver with robotics programmed with artificial intelligence is it increases safety on the tracks. There’s no individual who can be harmed performing high-risk exercises, decreasing the liability of automakers.
There are government regulations regarding the use of humans behind the wheel on proving grounds which can slow testing; these rules don’t come into play with an autonomous solution.
Humans require breaks and there are stiff safety regulations that dictate only one to three hours behind the wheel every twenty four hour period in many cases. On the other hand, robots can perform for an extended time period only needing to stop to refuel. And one of the most notable benefits of robotics is a more successful rate of repeatability that an individual just can’t match.
A Multi-Vehicle Control Platform
ASI’s Mobius for proving grounds is a multi-vehicle command and control platform that allows a single operator to oversee the operation of an entire fleet of automobiles. To accomplish this, vehicles are fitted with ASI’s Vehicle Automation Kit which includes our VCU (vehicle control unit) that communicates with Mobius.
The platform allows for the creation of custom paths to fit the parameters of desired tests, as vehicles can be programmed to accelerate, decelerate or perform any number of actions at precise locations. By fitting a beacon on non-autonomous vehicles, Mobius will recognize and account for them so both traditional and driverless vehicles can occupy a track at the same time.
Proven Results on Proving Grounds
Proving grounds using ASI’s Mobius technology see dramatically improved efficiency. One customer reported that tests with human drivers took 12 days to complete as a result of higher test failure rate due to human error, the restrictions of heavy safety regulations, and natural human fatigue requiring breaks. These same tests took robot drivers just five days to complete and logged high test ratings due to superior repeatability.
Car companies deploying ASI’s Mobius multi-vehicle autonomous technology for proving grounds enjoy faster, more accurate tests at reduced costs without putting a human driver in harm’s way.
Competitors’ solutions don’t allow the flexibility that Mobius provides in creating custom paths and events, which is why Ford, Toyota, FCA and Hyundai all rely on Mobius and ASI for durability and misuse testing. To learn more about the power and capability of Mobius, visit ASIrobots.com today!
The market for robots is booming and primed for further growth. Estimates by the International Federation of Robotics report unit sales of industrial robots grew 15% in 2015, while revenue increased 9% to $11bn.
Consultant firm ABI Research sees sales of industrial robots tripling by 2025. The popularity of robots is driven by their effectiveness in optimizing operations, lowering costs, and improving workplace safety.
Contrary to popular belief, automated machines don’t spell doom and gloom for the human workforce. The adoption of manufacturing robots in high-wage countries is allowing companies to reshore jobs, moving them back from low-wage countries like China, and employing locals back home.
Sportswear giant Adidas announced this year it will produce running shoes in a German factory staffed by robots and 160 new workers. It’s part of an overall trend where companies are investing in “collaborative robots” that are designed to operate alongside human workers.
Firms looking to incorporate automated vehicles into their workforce need look no further than ASI, a world leader in unmanned ground vehicle systems. We’ve developed a robotics platform that’s specifically designed for rapid integration into our partners’ ecosystems. Our robust technology ranges from driver assistance solutions to full, multi-vehicle autonomy.
We can easily automate any vehicle, regardless of manufacturer, using our Vehicle Automation Kit. It can also be leveraged for electronic, by-wire control of vehicles with controls capable of this function. Control functions are delivered via our Mobius software, the industry’s most advanced unmanned command and control system.
GLOBAL PARTNERSHIP PROGRAM
Through our Global Partnership Program, we’ve teamed up with OEMs and delivered automated solutions to leading brands including Ford Motor Company, Goodyear Tires, CNH Industrial, Boeing, General Dynamics, and Lockheed Martin. Fortune 500 companies and government entities trust us for reliable technologies that prioritize safety above all.
Our industry-leading engineers and researchers can serve as an extension of an existing robotics engineering team, or in lieu of one, in order to create and sustain a company’s competitive advantage. We offer three partnership models to choose from:
COLLABORATION – With innovations in global robotics developing at an exponential rate, stay ahead of the curve by leveraging our engineering. Whether it be for a single project or multi-year initiatives, we’re here to collaborate.
CONTRACT OUTSOURCING – If you have a disruptive idea, we can make it a reality with experts in business strategy, marketing, sales and creative finance to assist in creating liability isolated business entities.
EQUITY OWNERSHIP – Our aim is to find long-term strategic partners for equity positions. We offer creative and flexible ownership opportunities so our partners can secure access to our robotics solutions.
Whatever your current position, ASI is standing by, ready to help your organization reach its automation goals through a trusted partnership. To learn more about integrating ASI technology into your operation, visit ASIrobots.com.
A recent survey by the World Economic Forum found that 82% of executives in the mining industry plan to increase investments in digital technology within the next three years, and nearly a third (28%) expect those contributions to be significant.
This widespread commitment to data visualization, cybersecurity, robotics, and automation will have profound implications for what is a 402-billion-dollar industry.
It is very likely that organizations who move swiftly to adopt and implement these new technologies will emerge as market leaders with significantly reduced costs and overhead.
The same report found that early adopters may see an EBITA difference of nearly 70% compared to their laggard counterparts.
This stark contrast can be attributed to vast increases in operational efficiency. Companies that embrace robotics and automation will lower administrative costs, prolong equipment life, and be able to make better on the scene decisions.
There is also an increase in productivity associated with technologies like Robotic Process Automation (RPA).
RPA takes many of the repetitive tasks that are associated with a large knowledge workforce and mimic these activities through software to free up human counterparts for more creative and important work.
Mckinsey reports that one company saw a nearly 200% ROI from the RPA efforts, in the first year. Savings like this are leading many businesses to combine their IT and OT departments. These integrations will allow for unprecedented synergies in the field and at headquarters.
Over 23% of respondents in the Accenture Digital Technology in Mining Survey 2016 also reported that their organization has widespread adoption of robotic technology, and another 29% are in the pilot phase.
These means that physical robots will soon be an integral part of mining operations for the majority of industry leaders.
There will continue to be seismic shifts in the mining industry and many companies that are late to adopt will find themselves on the wrong end of a zero sum game.
To learn more about the future of mining, visit twitter and tag @AccentureMining
In our continued leadership innovating autonomous solutions for the mining industry, ASI has become a member of the Global Mining Standards and Guidelines Group (GMSG).
GMSG is an international mine operator-driven community that facilitates collaboration across the mining industry to solve common problems and develop standards, guidelines, and best practices.
As a corporate member, ASI has the opportunity to influence current and future initiatives, and shape the future of the global mining industry with respect to autonomy.
GMSG is presently developing an autonomous mining guideline for international use. Implementing autonomous solutions is a high-level priority for operators within the space, but due to the newness and general unfamiliarity with the technology, its integration has proven rather challenging for companies, in particular those that seek to scale up their current autonomous machines only to be frustrated by interoperability issues.
It is the group’s goal to share best practices for autonomous technology implementation in order to drive innovation, facilitate conversations with regional regulators, and optimize both mining safety and production.
For the purpose of creating a homogenous autonomous mining system, CMSG will borrow from the established work of outside industries in order to speed progress.
International protocols will guide manufacturers and tech providers in their innovation and development strategies.
It will also lead to consistency in output and process control so that the industry moves forward in lock-step.
And it’ll allow owners and operators to understand data requirements and standards, while contributing to the work of the International Organization of Standardization (ISO).
ASI is committed to engineering autonomous solutions that are interoperable with systems and equipment used all over the world.
As a member of the GMSG, we are in a prime position to establish the guidelines for autonomy, ensuring that our technology is standardized and able to provide maximum interoperability.
We are excited to take part in developing these global protocols.