The ultimate goal of automating business operations is to achieve a state where machines run the show, while you as a manager, control and direct flow of operations. The main tests of the effectiveness of automation are how often people must intervene in processes and how remote the supervisors can be.
At a high level, New Horizons, is an autonomous system that operates at a distance of 3 billion miles, and that has not had human intervention for nearly 10 years. I would contend that this is a major achievement in autonomous systems that we can learn a lot from to apply to infrastructure and operations automation.
Complexity: With New Horizons the design team realised that the variety and complexity of physical components were the most likely point of failure so they simplified the components (the antennae on a fixed mounting, fixed cameras etc). This improved reliability on the front line – but increased the complexity of the actions that the spacecraft had to take to achieve its goals. Getting a picture of Pluto back to Earth meant re-orienting the spacecraft, first for the picture with the correct camera, and then to line up the antennae with Earth to transmit. But all of these activities were soft; in that they were defined in the intelligent control systems on board and created no additional overhead to make more complex. In addition, they were defined in a way that could be improved in-flight by the domain experts.
Lesson Learned: Shift complexity into soft systems that can be managed directly by your subject matter experts. Not software systems that require third party software developers to create.
Agility: When the team were planning the journey they had some very questionable data about the trajectory of Pluto, in fact they had no idea if New Horizons would end up in the right place when they launched. They did have a rough idea of where Pluto should be in 10 years’ time, but they knew that they could give the spacecraft the general direction and that it is smart enough to autonomously maintain that chosen path.
New Horizons local operations could not wait for remote help when something out of the ordinary happened, so it had a myriad of local intelligent evaluation and decision rules that allowed it to deal with exceptional circumstances immediately and then maybe call for help to review if its strategic objectives should change. By analysing sensor information and known data, they could produce the operational intelligence required to modify the path for the spacecraft to follow without needing direct control of the processes.
Lesson Learned: Use hierarchical management and controls to maximise automation: Simple, local, short run local processes are best run closed loop with intelligent automated rules to manage outcomes and exceptions. Longer term objectives should be driven by processes that have both local situational awareness and operational intelligence. The higher up the hierarchy you can put human intervention, the less effort is required.
Visibility: To be able to develop new algorithms for controlling the spacecraft, the New Horizons team needed deep visibility into the sensor data and the actions, issues and exceptions that the spacecraft had endured. Visibility of past, present and future performance is critically important to the modification of the goals and objectives of the autonomous processes.
Lessons Learned: Combine and analyse real-time events, content, existing knowledge and known data to create situational awareness. Connect this information to modify local controls, as well as provide analytics for human a managers.
Defining Objectives: The New Horizons mission left Earth with an objective of flying by Pluto. The only fixed item was the resources it had to draw on. During its 10 year flight new techniques emerged which would allow better use of the physical resources on the spacecraft. The subject matter experts were able to update, re-train and re-program the spacecraft during those 10 years so that the best use of the physical resources could be made. All of this was not done before the launch, much of the flight plan, fly-by plan, sequencing and automation were loaded up to the spacecraft during the flight…ultimately resulting in a perfect fly-by and maximising the use of the equipment.
Lesson Learned: You do not need to automate everything in a process day 1. Get the objectives and operating principles right and you can evolve the perfect plan during flight – so long as your subject matter experts have direct access to the automation platform.
You will not be surprised to learn that Cortex Intelligent Orchestration platform enables these and many other lessons learned from the space industry. In fact, Cortex shares some core technology and techniques in expert systems with the space industry.
If you are looking to create autonomous operations in your business, Cortex is the enabling technology of choice for many large organisations.