The pace at which we are developing artificial intelligence (AI) has been rapid over the last few years. From helping to keep humans from harm’s way with advancements in autonomous vehicles through to looking at how we reduce the volume of repetitive tasks that we do, the evidence of intelligent machines starting to perform tasks by themselves is clear to see.
Autonomous navigation systems
Autonomous navigation systems are playing an increasingly important role in helping provide safer and more efficient mission management and path planning – reducing human error and costs whether on land, sea or in the air. This software can be used effectively to provide the successful achievement of three critical functions:
Discover: Survey; inspect; surveillance
Deliver: People; supplies; effect
Plan: Safe and secure autonomous operations
tpgroup‘s unique Northstar (N*) path planning software
Our cutting edge patent pending N* software produces dynamic risk aware optimised path planning for autonomous platforms – inspired by the academic technique of ant colony optimisation. The key elements are as follows:
Constraints-based path planner
This innovative software adds value to any autonomous system or capability by generating optimal routes against multiple dynamic user defined constraints of specific vehicle characteristics, performance, fuel efficiency, time and cost.
It factors in the ever changing environmental elements, from forecasts or detected by on-board instruments, combined with input from mapping – to avoid any moving or stationary, mapped or unmapped obstacles.
Digital world building
At the heart of our autonomy capability is digital world building. For instance in the maritime domain, any known data (such as maps, weather forecasts, and tidal streams) is fused together to create a synthetic environment.
Further inputs from on-board instruments – such as radar, LiDAR, sonar bathymetry, stereo vision, depth and cameras – are layered on top to give additional dimensions and a more rounded view of the surrounding environment. This produces an optimal path that it is possible for the vessel to navigate through.
Future application across multiple sectors
Our N* software can easily be utilised for other vehicle platforms across land, airborne, space and sub-surface applications within the commercial, defence, and emergency response domains.
This holistic planning approach also enables the simultaneous planning of multiple robots – de-conflicting and optimising their paths as a team. This enables collaborative exploration of the environment, for example in a disaster response mission or other hazardous situations such as offshore wind farms, oil rigs or a nuclear setting.
We have developed an autonomous navigation system to enable a safer maritime industry – reducing human error and the level of manpower normally required to operate a vessel.
One of our current projects is supporting a MoD autonomy project to make mine retrieval safer and been testing the software and its responses to the environment in a series of sea trials.
Keeping humans from harm’s way whist achieve greater operational agility is driving the development of our autonomous navigations systems for multi platforms on land. We see this as a critical development; for example, in helping emergency services to deal with natural disasters and getting much needed supplies to remote and war torn parts of our planet.
There is also a huge opportunity within the defence sector and we are working with the US Department of Defense to develop the software for convoys of land-based vehicles.
Autonomous navigation FAQs
In artificial intelligence (AI), autonomy is no different. An AI system, also known as a robotic operating system, can carry out its own processes and operations without external control.
Autonomy requires that the system be able to do the following:
• Sense the environment and keep track of the system’s current state and location
• Perceive and understand various data sources available
• Determine what action to take next and make a plan
• Take action only when it is safe to do so, avoiding situations that threaten people’s safety, property or the autonomous system itself
Autonomous operating systems vary from simple mass marketed robot floor cleaners to complex Defence equipment.
Whilst there’s industry talk about autonomous systems being used, the vast majority are semi-autonomous. For example, cars with driver support systems such as lane keep assist and advanced braking systems are semi-autonomous, as are robot vacuums and most unmanned aerial vehicles (UAVs and drones). Most fully autonomous systems are currently too costly (data intensive, power consuming and budgetary) or unsafe for widespread public use.
It wasn’t until 1940s that a programmable digital computer was invented – the Atanasoff Berry Computer. This inspired the idea of creating an “electronic brain”. In the 1950s, Alan Turing proposed a test that measured a machine’s ability to replicate human actions to a degree that was indistinguishable from human behaviour – and the decade also saw the field of AI research founded. The acceleration of knowledge has grown with every decade to the tangible reality that we see today – from robotic hoovers to the future of travel and work.
AI will help keep people safer in removing human error – from changing how we navigate and use our oceans to using robots in dangerous work. It’s also recognised that many jobs haven’t changed since the last industrial revolution – there remains a huge amount of repetition in roles or ‘dirty’ tasks that could be better carried out by robots. For example, data centric tasks to a high level of accuracy can be completed using AI in a matter of seconds – compared to hours for humans.
Businesses can benefit from the operational agility that AI brings from increasing operational flexibility through to greater efficiency and productivity. Further there is the opportunity to create new types of roles to bring greater employee satisfaction and harness previously untapped creativity.
There are a number of different autonomous navigation technology in use. Some use remote navigation aids in path planning while, for other forms of autonomous route planners, the only information available to compute a path is based on input from sensors aboard the platform.
Once the basic position information is gathered in the form of triangulated signals or environmental understanding, the autonomous navigation robot has to apply machine learning to translate why it is leaving its present position into a route and motion plan. As part of this, it might also have to take into account the intentions of other autonomous robots in order to prevent collisions, as well as constantly being aware of hazards in its environment. Most of these systems will revert back to their original path once obstacles are overcome.
tpgroup has advanced the more traditional autonomous navigation systems in the market – producing risk aware optimised path planning for autonomous platforms that responds in real-time. This is inspired by the academic technique of ant colony optimisation.
In the UK, four institutions are at the centre of the UK’s AI strategy. This includes the Alan Turing Institute, the Office for Artificial Intelligence, the Centre for Data Ethics and Innovation, and the AI Council.