Everything that moves
If you’re developing Embodied AI products for Automotive, Robotics and other industries, Kognic is designed for your business. With our heritage in powering perception systems for mobility, we’ve built a proven toolset that addresses all facets of dataset engineering.
Since 2018, Kognic has been providing our sophisticated platform to speed development, deployment and testing of ADAS / AD solutions across OEM, Tier 1, Tier 2 and beyond.
With the Kognic Platform, you get full coverage from dataset curation, to expressing your use cases, creating the right guidelines, through annotation and measuring the performance of your system. Kognic guides you through the whole journey, ensuring that you deliver trusted and high-performing driving experiences.
Future-proofing ADAS / AD
“When you start to know the Kognic team, even if it's at a small scale, you build up trust”
As one of our earliest and ongoing relationships, Kognic has worked closely with Zenseact to create and manage efficient dataset shaping for automated and autonomous mobility. By using advanced sensor fusion and our Refinement tool's Loss-Ranking view, we enable Zenseact to meet their ADAS performance criteria demanded by the SOP for Volvo Cars.
Enable an effective way to conduct quality assurance while iterating datasets.
Our Refinement capability enabled Zenseact to get actionable insights by comparing model predictions to their datasets to focus subsequent QA efforts.
Higher dataset quality and new insights about their model behavior resulting in 28% reduction in QA investment.
Increasing the impact of end-to-end ground truth
Multiple missions, global teams and demanding use case coverage
Kognic is proud to work with Bosch and their global ADAS/AD product teams across the full breadth of our platform. In a combined agreement to provide both software and ground-truth deliveries, Kognic is addressing a large roster of advanced use case complexity for Bosch’s commitment to bring their ADAS/AD products to market for multiple OEMs.
Development cycles, from data ingest to model output, had grown too long and were risking OEM release calendars
Shorten development iterations through integrated dev-to-prod annotation workflows, and reduce production cost through pre-annotation and broader automation
Kognic's unified toolset working across the entire lifecycle of data production and deployment has reduced overhead and increased uptime.
Accelerating engineering precision
“Kognic has a very talented team who understands the problems of automated vehicles' safety, sensor fusion, and large-scale data processing. Their platform, scalability and quality make them our preferred partner.”
For our ongoing relationship with Qualcomm, Kognic provides both best practice and customized workflows to their globally distributed teams. Their promises to OEM partners, to deliver ADAS/ AD systems, are made real by Kognic’s bullet-proof ability to deliver under pressure and under tight deadlines.
Sub-optimal results from previous software vendor, including lack of automation and workforce mgt. tools, led to missed expectations and delivery.
Provide a flexible mixture of software and perception expertise to accelerate annotation, optimize dataset engineering workflows and regain program momentum.
Stabilized data supply by delivering on promises, both in regards of quality and time, leading to increase in models sent to testing.
Aligning guidelines to set a solid foundation
“Some companies are mainly focused on the end result, which is basically: we have the annotators, it will take this amount of time, this is our price and that's it. Kognic goes beyond and really guided us in the initial steps, and that helped us stay the course and achieve our initial goal.”
For ZKW, a leading supplier now part of LG, Kognic focused its program and platform efforts on establishing solid guidelines for their perception system development. Using Kognic’s Guideline Agreement Process, ZKW product owners were guided to deconstruct their use cases, together with Kognic perception experts, to reveal complexity-to-value. Setting up the guidelines wasn’t only helpful to make sure the plan didn’t deviate from the initial goal but also to make sure the focus was on getting the right data from the start.
Gain early and detailed understanding of ADAS / AD use case variety and their impact to decisions about data composition.
Kognic choreographed our Guidelines framework to build acceptance criteria into software workflows.
Reduced model performance errors by aligning dataset composition with use case objectives.
Everything that moves will need perception to advance further automation and autonomy. In the robotics ecosystem, the unique needs of fine-tuning datasets for supply chain and logistics environments is rapidly increasing.
Shaping datasets for warehouse dynamics, such as inventory handling – “pick & pack”, environment sensing, quality control and even last – mile delivery, pose new challenges for Embodied AI. While controlled environments are common, the unique actors within these experiences bring new complexities to dataset engineering.
Manufacturing – Heavy industry
With the growth and impact of predictive maintenance as a key workflow for heavy industry, live models require dependable, high quality and relevant data at scale.
Kognic’s Platform tooling, driven by its core human feedback and leading annotation engine, helps OEMs fine-tune these models for prediction and prevention of critical system events.