Automated development tool chain for embedded systems
91% less integration time
Automated development tool chain for embedded systems
Key use cases
An advanced equipment manufacturer reduced both R&D costs and time to market by 15% by running virtual reality-enabled hackathons
Case example 2 / 8
The company made a breakthrough with its next-generation model of a large stationary electronic device by developing new product versions as 3-D models in virtual reality hackathons. It brought people from various departments and locations together in the same room, either physically or virtually, to push a virtual reality prototype through multiple cycles of review and adjustment, making refinements in real time. This accelerated the design process and made it possible to solve problems quickly in an agile manner. With the right tools in place, the cross-functional teams were able to alter the prototypes even more quickly and estimate the cost implications of potential design improvements in real time.
This resulted in a reduction in time to market of 15 percent.
High tech player reduced integration time more than 90% with automation
Case example 1 / 8
A high tech player introduced a highly automated software development tool chain to cover the R&D process from customer change request to release. The solution uses automated test runs on the target hardware (an embedded system), including the flashing of firmware and an easy-to-use web-based workflow. The decentralized architectural layouts ensure global, scalable performance and the total tool time is approximately 30 minutes across all target hardware categories. Worldwide, the company makes over 10 million deployments a day.
The result is a faster and more agile development process with a 91% reduced integration time from the code freeze to the successful compilation of the target build. Additionally, the firm lowered the defect density via strict quality gates and achieved a 50 percent increase in mean time between failures.
–91%
duration of code freeze to successful compilation of target build
Virtual reality-enabled hackathon
Virtual reality-enabled hackathon
15% faster time to market
Digital product twin creation
Digital product twin creation
12% EBIT increase
R&D project efficiency analysis
R&D project efficiency analysis
11% less engineering and design costs
Usage-based feature optimization
Usage-based feature optimization
Elimination of free and unused features
Digital should-costing
Digital should-costing
Benchmarking of >100,000 data points
Paperless work instructions
Paperless work instructions
+10% production yield
Software product cost optimization
Software product cost optimization
40% software cost reduction
Automated assignment of defects to
solution domains
Automated assignment of defects to
solution domains
11% better accuracy of defect assignments
A large global automotive OEM gained 12% EBIT through the introduction of digital product twins
Case example 3 / 8
High product and portfolio complexity, an increased share of software in products, and an ever- increasing globalized world pose challenges for global automotive OEMs. Simultaneously, trends in digitization, virtual products and technical compliance are shaping the industry more than ever. A global automotive OEM has introduced digital twins for each conceptualized, buildable, and built product. It made a digital replica of each product containing all data for each component and variant available throughout the entire product lifecycle.
It achieved an overall earnings before interest and taxes (EBIT) impact of over 12 percent based on both revenue increases and cost savings. The cost savings resulted from increased transparency, twin availability throughout the entire product lifecycle, enhanced product-specific information and the continuous product updates a digital twin makes possible.
An automaker realized 11% savings in engineering and design by analyzing indicators for R&D project success
Case example 4 / 8
An automotive company analyzed historic data for all its R&D projects over the past 10 years using machine learning. The analysis included product property profiles, project and milestone planning, bill of material (BOM) models, change requests, quality reports, vehicle usage reports and employee communication data. The model describes project efficiency by output parameters like hardware cost, personnel cost, and usage, and tests the influence of over 20 input parameters like the amount of communication among team members, number of iterations, number of child projects and number of change requests on these efficiency indicators. This analysis helped the automotive company predict future productivity as an “early warning” tool for targeted interventions.
This effort improved its R&D efficiency by reducing time-to-market and development costs while guaranteeing superior product quality and customer-valued product properties.
–11%
engineering and design costs
An OEM captured EUR 20 million in lifetime savings using quantitative decision support tools to cut unused car features by evaluating real customer vehicle data
Case example 5 / 8
The company used vehicle sensor data of customer behavior and patterns on specific features to focus their R&D effort by eliminating dispensable car features. It tracked CD/DVD player use via central car data, acceleration behavior from powertrain sensors and temperature profiles from different regions for cooling/heating. The model helped it make the trade-offs involved in adding additional content versus capturing additional price points. It provides recommendations on product design-driven material cost optimization based on real customer needs.
The OEM, for instance, achieved EUR 20 million in lifetime savings by eliminating the CD/DVD player as a standard configuration in the US market, based on the finding that 98% of car owners use the CD/DVD less than 15% of the time.
Elimination of unused features
An automaker used an algorithm based on competitor data to reduce its manufacturing costs, adjust pricing, and simplify its own offering structure
Case example 6 / 8
The company wanted to maintain competitive customer value by strategically positioning its own car offering in a complex and dynamic competitive landscape. Faced with finding the right balance between spending and material cost at the subcomponent level, the company developed an algorithm based on its own and competitor data.
The algorithm uses the company’s own bill of materials (BOM) of relevant components and manufacturing costs, along with information on competitor cars from price lists, used-car data, and car magazines and ended up with over 100,000 data points. It analyzed and visualized its gap to competitors in material costs and customer value ratings. The model also provided strategic insights and recommendations for an optimized offering structure.
>100,000
data points
A global automotive tier 1 supplier sought short-term software cost reduction opportunities across a wide portfolio of automotive products
Case example 8 / 8
A global automotive tier 1 used Softcost diagnostics to identify a 40 percent software cost reduction. The algorithm finds ways to reduce project complexity and assess vendor costs, correlating them to find best fit in terms of vendors and outlays. It also calculates the design complexity of multiple implementation options and analyses fact-based trade-off simulations to explore scheduling and resource and cost requirements for each scenario.
The tool validates results using industry benchmarks from the analytics tool software database.
–40%
software cost
A car company used an algorithm-based assignment of defects to solutions domains
Case example 7 / 8
Allocating defects to the right domain as fast as possible can speed up development processes. A global OEM had to assess over 250 defects each day and send them to the right solution domains (department) for solving. Wrong allocations would lead to reassessments of the defects and take longer to resolve due to waiting times at the wrong solution domains.
The company developed a machine learning algorithm based on over 15 tested models and 100,000 historic defects for training that improved its first-time right allocation from 55 percent (manual by human) to 61 percent (algorithm) and thereby reduced the average defect solving time. Besides freeing up full-time employees for value-creating work, allocation happens in almost real time. The automaker achieved this improvement to 61 percent at the beginning of the project and going forward plans to improve accuracy up to 80 percent.
+11%
accuracy of defect assignment
Jabil Circuit gains 10% production yield through interactive work instructions
Case example 9 / 9
Jabil Circuit uses tablet-based interactive work instructions from Tulip to guide electronics assembly and run process analytics at the same time. Customized apps provide the operators with visual instructions about the assembly steps, enabling them to track and report quality issues in real time. The system has fully replaced former paper-based work instructions and audit procedures, and the data gathered from the shop floor are continuously analyzed by Tulip’s analytics engine. Within four weeks of implementing the digital work instructions, Jabil Circuit gained a more than 10% increase in production yield and a 60% reduction of manual-assembly-related quality issues.
+10%
production yield
+50%
mean time between failures
10mn+
deployments per day
–15%
time to market
+12%
EBIT
Benchmarking of