Case Study: System Optimization and Advanced Debugging for Tevel Aerobotics Technologies
A Collaboration in Embedded Vision and Yocto Platform Development
Tevel Aerobotics Technologies, a pioneering company in autonomous fruit-picking robotics, combines aerial robotics, computer vision, and AI-driven decision-making to automate agricultural harvesting.
TandemG partnered with Tevel as an R&D extension team, focusing on strengthening the reliability, scalability, and maintainability of the embedded software platform that powers Tevel’s robotic systems operating in complex, real-world agricultural environments.
Challenge
During development of Tevel’s embedded vision system, the engineering team encountered a persistent software malfunction within a third-party image-processing library.
The core function repeatedly returned a failure code (-1), even when the image was successfully processed.
This inconsistent behavior disrupted the image-analysis pipeline and limited system stability.
Despite extensive efforts by Tevel’s internal engineers, the root cause remained undetected for months, delaying further integration and field testing.
Solution
TandemG’s engineering team executed a deep debugging and root-cause analysis within the Yocto-based environment, utilizing advanced tracing tools, detailed runtime logs real time, and custom instrumentation.
By closely analyzing memory allocation patterns and asynchronous API calls, the team identified a logic flaw in the library’s error-handling mechanism that caused false-negative return codes.
After isolating the fault, TandemG implemented a corrective patch and verified the fix across multiple hardware configurations and test cycles.
This solution restored stable image processing and eliminated a critical reliability bottleneck that had hindered system progress.
Yocto Platform Upgrade
In parallel, TandemG led Tevel’s Yocto platform upgrade from version 3.0 (Zeus) to 3.3 (Hardknott), including:
- Integration of Docker (native), K3S (v1.22.5), Fluentbit, and Chrony (v3.5)
- Implementation of full NetBoot capability for root and boot partitions to support future OTA (Over-The-Air) updates
- Custom Device Tree (DTS) modifications tailored to Tevel’s proprietary hardware board
- Validation and verification on both EVM prototypes and production-level hardware
The upgrade significantly improved the system’s modularity, maintainability, and operational stability – creating a robust foundation for future product scaling and deployment.
Technology Overview
- Platform: Yocto Linux (Hardknott)
- Containerization: Docker, K3S
- Monitoring: Fluentbit, Chrony
- Programming Languages: C / C++
- DevOps & CI/CD: GitLab, Jira Cloud
- Debug Tools: GDB, Memory Tracing, Custom Instrumentation
Results & Impact
- Root cause of a long-standing image-processing malfunction successfully identified and fixed
- Consistent and stable image-processing pipeline
- Upgraded Yocto platform with enhanced modularity and FOTA-readiness
- Faster debugging cycles and improved reliability
- Strengthened foundation for production and scaling of Tevel’s robotic systems
Key Highlights
- Deep debugging expertise resolving a complex third-party library issue
- Yocto OS upgrade from Zeus to Hardknott
- Integration of Docker, K3S, and Fluentbit for modular, monitored environments
- Custom DTS and NetBoot support for production scalability
- Proven reliability improvement validated on production hardware
How can an external team resolve an issue that stalled internal development for months?
Our advantage is focus. We brought in Deep Debugging experts with specialized tools for deep analysis. This freed Tevel’s core team to continue product development, while we pinpointed the hidden flaw in the third-party library.
Why was a Yocto upgrade necessary, rather than just fixing the bug?
To future-proof the system. The upgrade enabled the integration of Docker and Over-The-Air (OTA) update infrastructure. While fixing the bug solved an immediate problem, the upgrade made the system production-ready.
What makes this solution specifically tailored for scale?
The combination of operating system stability and container modularity (K3S). This architecture allows Tevel to manage and update entire fleets of robots easily and securely, rather than just individual units.
In Summary
TandemG is an R&D extension company specializing in full-stack product development – from embedded systems and IoT to cloud platforms and UX design.
We partner with startups and enterprises to deliver reliable, production-ready technology solutions that scale.
Want to talk? Contact us