Mechatronics and embedded software add complexity to smart device design

The term 'mechatronics' highlights the importance of systems that combine mechanical parts, actuators, and hybrid electronic/computer-based controls. Engineers and designers working on a range of products from consumer goods to industrial machines have embraced mechatronics, and developed concepts, skills and techniques to deliver more flexibility and performance at lower cost. However, this has not been easy. Bringing together multiple technical disciplines exposes issues including organization structures, standards, and complexity.

Practitioners have at every step been pushing the boundaries of the tools they use, tools that very often have their roots in a single technology – mechanical CAD, electronic design automation, software development, stress analysis, kinematic analysis and so on. So not only has it been necessary to solve problems of integration of the various technologies, but, in parallel, designers and engineers have had to solve problems of integration of the tools they use. Systems-engineering approaches have helped structure a suitably broad view - the product as well as its development and operating environments.

Today, the scope of the mechatronics problem and opportunity has grown dramatically because of the low cost and easy availability of network connectivity. Every device has potential to be connected to the network, so designers have to consider the relative merits of all possible divisions of function between the 'local' device and 'remote' capabilities accessed via the network.

There are examples in every sector - more products that phone home for software updates, medical imaging devices that use a connected service for archiving, jet engines that transmit sensor readings to service desks the other side of the world for in-flight monitoring and analysis, consumer electronics devices that depend on remote sources of music and images, agricultural machines that regulate fertilizer concentrations according to GPS location and historical records of yield, and many more. In all these cases, design choices about which side of the network interface boundary will provide the resources that support the functionality are critical. These choices impact in-service performance, and indeed the value proposition for the product and service.

The network can also enable innovation of a type that can transform the competitive playing field. For example, industrial machines often come complete with controllers that include a network interface and embedded software that will deliver machine status information and accept control parameters via a web-page – very useful for remote monitoring and control.

So it is not a big step for a machine designer to add more embedded software into the controller. Imagine adding software to integrate the machine to the customer's corporate asset management system. Delivered with professional services, this capability could allow a customer to rethink and perhaps automate more of their maintenance activity.

Traditional interval based servicing can be replaced by as-needed adjustment. A machine builder could use this embedded software to change the basis of competition – in addition to the established machine requirements such as precision, power consumption, and flexibility, the new capability offers a new way to reduce whole-life costs. And this kind of thinking can be repeated in every product that can carry a network connection.

Even staying focused 'inside' the product, embedded software is the technology that will provide the innovation in the next generation of products. Touch screens, cameras, microphones, GPS, motion sensors and so on are all increasingly available as 'commodity' components. These offer 'standard' performance that comes alive when designers find new ways of assembling them with embedded software to coordinate their functions.

In addition to these 'static' components, equivalent development of actuators signal a future in which machines of all types will be built from low-cost standard mechatronics subsystems assembled around a platform architecture or framework, and integrated by the embedded software loaded into the platform. Individual product lines will depend on the scope of the platform; differentiation will depend on the capability of the embedded software. Indeed, the automotive sector has for many years been moving towards this structure.

With software taking this central role, both inside products and for products connected to networks, it is not surprising that across all the industries that create 'smart' products, engineering management teams have a sharp interest in the way their companies will handle this critical technology. They know they want to manage every technology that differentiates their products from the competition – and this now includes software.

But software development can be a bit of a culture shock. The widespread visibility of 'apps' in the consumer world has helped de-mystify the intangible nature of executable files. But the relationship of these to complex build structures of source code, the subtle effects of interaction of multiple software objects around a network, the threat of malware, and the willingness of software engineers to handle changes at rates ten or a hundred times the rates you might expect for mechanical components are all factors that take a bit of getting used to.

These engineering management teams need a method or process that gives them the visibility they need, and provides a framework within which multi-disciplinary team members can be creative, avoid being constrained by artificial boundaries between each technology discipline, and assemble toolsets to meet the needs of the range of disciplines they handle.