By Wolfgang Patelay

The internet is playing an important role for another software-centric industry " industrial control and manufacturing. In the manufacturing industry IPC and PLCs are controlling most of the processes. Therefore at the factory floor there are working nowadays computer controlled and therefore software-centric systems. It starts in the storage where automated systems deliver goods via automated transport systems to the working station where they have to be further processed. At the workstation maybe robots " which are software-controlled - take these parts and build for example a car body. The resulting primary product is also computer-controlled automatically checked whether the weld seam is correct. The software which is used to do the job at the workstation is stored locally on the machine respective its controller. This decentralized software storage and use is standard since the advent of the personal computer. But with the omnipresent Internet these little local software-centres are becoming more and more cross-linked.

This world-wide computer link enables now totally new opportunities: the so called "cloud-computing". With cloud-computing experts are describing a new web-centric way of computing which shifts programs, data and computing power from the PC into the web. The user has access to the power of computing centres and can use always actual and virus-free software. The door to the "cloud" opens an Internet-browser like the new Google browser Chrome which enables the easy and quick access to the power of the "cloud". This new way of computing would avoid the cost of maintenance and service for locally installed hard- und software. Experts are guessing that to date up to 80 per cent of the total system cost is spend for maintenance. Saving this money is a great advantage especially for large organisations.

By Wolfgang Patelay

But this also means the machine, in this case the car, has to be equipped with many sensors and actors to see, hear, feel, smell und react accordingly. In modern cars driver assistant systems support the driver in many ways. Optical sensors measure the distance to the cars in front and adjust the appropriate distance in relation to the actual speed. Lane departure warning systems recognise unintended leaving of the lane and create a warning to the driver. Anti-collision systems evaluate the traffic around the car and react accordingly to avoid a collision. And there are many more examples like anti-lock brake systems, traction control systems, night vision systems and so on which proofs that cars becomes smarter.

And even on the manufacturing floor youll find smart machines. If you look at a modern electronic production line there are many intelligent machines combined to work together. Automated storage systems feeding conveyors which transport the PCBs to assemble to the pick-an-place machines which contain intelligent feeders. These feeders "know" exactly when and in which order the pick-and-place machine needs the components to be placed. The following production step soldering is also accomplished by smart soldering machines which know which board has to be soldered and use the specific soldering profile. And at the end of the production line smart test systems checking the circuit function and analyse possible failures. Due to their intelligence there are only few operators needed to run these lines.

And there is another excellent example of smart machines - robots! Experts forecast a bright future for these autonomous machines. Nowadays robots are integrated into production lines to do precisely hard welding tasks but in future these robots will work close together with human colleagues, will do the housekeeping, support elder or handicapped people during daily life, and in a few years it is planned to do a soccer game between a human team and a robot team. This means there must be extremely smart machines to carry out all his tasks. And it also needs well educated and skilled people to develop, build and use all this smart machines.

By Wolfgang Patelay

But human beings interact with other human beings via language and not via any kind of interface. To enable humans to interact with computer controlled systems via their language researchers and specialists are working since approx. 30 years to create speech recognition systems. And they did succeed: nowadays speech recognition systems are matured and used in various applications. A good example is the navigation system in a car which is more and more controlled by the voice of the driver - other functions like telephone and radio as well. Speech control of these functions increases the security dramatically because the driver can concentrate on the traffic - he or she has not to look at the navigation systems and press buttons or touch the screen to operate it.

Navigation systems are by far not the only application in which speech control is used nowadays. You can buy mobile phones which does not require any longer to dial the number " you talk to the phone and advice it the name of the person you want to call and youll be connected. Homes for seniors or disabled persons are equipped with speech control to operate for example shutters, TV, and telephone. Surgeons are able to control the light intensity in the operating theatre by voice. If you call civil services, banks, or insurances youll realise that there are more and more computer voices to answer your questions or requests. In Germany customers of Postbank are able to remit money, ask for the position of account and so on by a phone call. Pilots of the Eurofighter can talk to their "bird" to give it instructions. To write text its not necessary any longer to type in the text into a computer " via speech recognition software you can dictate it to the machine.

By Wolfgang Patelay

And youll find a comparable situation in electronic systems for industrial applications. There are PLCs and IPCs with standardized hardware for the dedicated application and - the software makes the difference, too. The more complex an application in the industry is the more complex the software has to be. This means a big effort to develop, verify, and debug the software. Although there are powerful tools commercially available to support the software designer to develop it is nearly impossible to design error-free software. Consequently the verification and debugging of software requires more up to 70 percent of the whole development process. And the situation becomes even worst because the complexity of software will increase dramatically further in future due to the integration of industrial systems into networks which brings security aspect even to embedded software.

If you could reduce the effort of debugging software - which means you have to develop error-free software - you would safe cost. Consequently the quality of software plays an increasingly important role in all segments of the industry. Due to the IDC market survey "Improving Software Quality to Drive Business Agility" spending enterprises up to 14 Million Euros (about $ 22 Million) to debug and correct software errors. The emerging multi-core designs make this situation worse because the complexity and therefore the effort of verification and debugging will remain at very high levels. The problems of poor software quality are caused by various factors like increasing code complexity, distributed software design teams, outsourcing, out-dated code, open-source code, and multi-thread applications. Half of the attendees of the survey admitted that in the first year after a software release up to ten critical errors occurs which requires patches. The expenditure of time to fix the bugs was between on day (16%), two to ten days (66%) and eleven to thirty days (11%). Due to the estimation of the attendees of the survey avoiding the software bugs would result in saving about 32 percent of the entire software development cost.