Connecting the Latest Embedded Vision SystemsFollow article
I love science fiction. I may not be as passionate about it as some, but it has still formed a large part of my life. If proof were needed, I am wearing a t-shirt depicting the robot Bender from the animated TV series Futurama as I type this.
For me, robots are one of the elements of science fiction that I enjoy, and there is plenty of inspiration to choose from. After all, artificial beings have formed a central theme in science fiction since the very first stories well over a century ago. The list of robots represented in science fiction is enormous. Even if you are not a science fiction fan, you could probably think of perhaps a dozen fictional robots, just from your own memory.
Have a go. I’ll wait.
I expect that your list includes C-3PO and R2-D2 from the Star Wars universe, and maybe BB-8 as well. It might include Data from Star Trek or even the Robot from Lost In Space. Robots in fiction have been used to represent almost every type of character – a figure of fun, a saving angel or an arch-nemesis. When portrayed in fiction, the eyes of the robot are of key importance. When describing a robot, its eyes can be used to highlight its humanity – or lack of it.
When robots actually entered our lives in the 1960s, they were nothing like those promised to us by the likes of Isaac Asimov or the studios of Hollywood. They were dumb machines, designed to perform repetitive tasks with precision and had little in common with those of science fiction.
As the electronics industry has matured, and the sophistication of machines has advanced, the design of robots is moving closer to the promise of science fiction. While we are still some way from truly intelligent machines capable of independent thought, modern robots exhibit remarkable performance. In order to function in the real world, however, robots need to gather accurate information from the environment around them.
Applications for Embedded Vision
Embedded vision systems provide the eyes for the current generation of robots. Whether using purely video input, or combined with other sensors including radar and lidar (light direction and ranging), the power of modern computing uses these systems to measure distance, identify obstacles, and navigate through the potentially hazardous factory environment.
There are other industrial applications for embedded vision systems, even in the factory. One area of interest is that of machine safety. Maintaining a safe working environment continues to grow in complexity as manufacturing plants become more flexible. Machines must detect possible hazards, and interpret them in real-time so that the right action can be taken to prevent accidents. It is only with the sophistication of modern embedded vision systems that this has become truly possible.
Embedded vision systems are also vital to efficient manufacturing. They are able to replace or enhance manual inspection as a key part of quality control. Traditionally, manufacturers were forced to choose between random sampling or 100% inspection.
Modern embedded vision systems allow automated, 100% visual inspection. Each item can be imaged by a dedicated camera and its dimensions compared to a standard in real-time. The result is that faulty products can be identified immediately and quarantined. Inspection in real-time also provides important data that can be used to understand maintenance needs and even predict future failures.
Embedded vision has found applications outside the factory too. The explosion in the field of autonomous, self-driving, cars has been in the headlines recently, and embedded vision systems will be crucial to the safe operation of these vehicles in the future.
Connecting Vision Systems
This profusion of vision systems – cameras and their supporting hardware – comes at the same time as the continued miniaturization of electronic devices. For many years, coaxial cable was used for video connectivity, but with the increased use of digital outputs, other solutions have come to the fore.
An ideal solution for compact, high-performance camera systems is the use of flexible printed circuits. These have traditionally used a polyimide-based substrate and printed using similar methods to conventional PCBs. A copper layer is bonded to the substrate to provide conductivity, which is then masked and chemically etched. This removes the copper from the un-masked areas to leave the final circuit, to which components can be soldered using conventional techniques.
This subtractive process is expensive due to the inefficient use of materials and the relatively high cost of the polyimide foil. The unusual shapes that are frequently required in FPC applications are even more expensive as the result of poor material utilization.
As an alternative to polyimide, engineers have been experimenting with a number of different materials. Recent advances made with conductive inks have allowed designers to use silver printing on polyester instead of copper FPCs. The construction of circuits with conductive inks is far more efficient than traditional etched copper techniques. These inks now permit manufacturers to create traces as narrow as 0.005” (0.127mm) with the spacing of a similar size, enabling these printed circuits to compete with traditional high-density PCBs.
LEDs, resistors, capacitors and an integrated circuit bonded to a silver printed circuit
Alongside the use of conductive inks, improved bonding technology is creating other opportunities. One of the problems with flexible circuits is the bonding of components onto the substrate. Established techniques used conductive epoxies to anchor components limit the ability to attach fine-pitch devices onto the substrate. This has been overcome with new bonding materials that allow the attachment of complex devices to flexible circuits. Microprocessors and other components, including the latest generation of cameras, can now form part of the flexible circuit.
This new manufacturing process is flexible enough to allow multi-layered circuits to be created. Solvent-based inks can cure in just a few minutes at 200° to 300°F and can be over-printed with a dielectric layer to provide electrical isolation and environmental protection. Using this method, it is possible to add further layers to create a more complex circuit.
Embedded vision devices are vital to the latest autonomous machinery, both inside the industrial environment and beyond. The demands for systems that are both smaller and yet more capable will continue to grow, forcing engineers to find new and innovative ways of connecting cameras. Technologies like flexible printed circuits and combining with developments in materials to provide designers with fantastic solutions. Could they be right for you, too?