Antenna design with a twist of AI gets great results faster

Here, Aitor Moreno, Cloud Product Manager, Ignion, explores how AI-driven design tools like Oxion can simplify and accelerate the complex process of antenna selection and integration in wireless product development. He discusses how these tools empower engineers, even those without specialised RF expertise, to optimise antenna performance and achieve rapid, accurate results.

Wireless is especially suitable for people and assets that are on the move, as well as for communicating over long distances and extending networks into places where wired infrastructure cannot be installed.

Organisations that do not have specialised wireless knowledge in-house tend to use plug-and-play wireless solutions or other simplified approaches to complete these aspects of their new product designs or may rely on external consultants. However, an approach that helps innovators accomplish more of the RF design independently can give more control over the development schedule and the performance and features of the finished product.

Antenna design challenges

When designing new wireless products, there are always complex challenges surrounding antenna selection and design-in. The system performance is usually extremely sensitive to the characteristics of the antenna as well as any connecting wires or PCB traces. The selection of balancing and filtering components that connect the antenna to the RF module is also critically important. Moreover, designers need to consider the antenna position, including whether it is mounted on the circuit board or off-board, as well as the proximity of other system components, such as a battery or the enclosure.

The influence of any other antennas within the system must also be considered. Any board-mounted antenna usually requires a clearance area in which no components or circuit traces are allowed to minimise unwanted interferences. In addition, if a system such as an asset-tracking device uses multiple antennas, such as GNSS for precise location and a cellular or LPWAN antenna for the IoT connection, a minimum separation distance is required to prevent coupling effects.

Antenna selection, interconnect, and positioning critically affect the performance of the entire system and can influence whether the desired communication range and coverage can be achieved. The RF-system performance can influence the battery runtime, if excessive transmitter power is needed or if data packets need to be frequently resent. Often, size and layout constraints restrict the design team’s choices and if the operating environment is complex or the design must manage different frequency bands, then achieving the desired performance can become very tricky.

Finding a suitable antenna can require designers to assess many potential candidates. Manually inspecting datasheets can be time consuming and different product manufacturers often express their products’ data in different ways thereby making accurate comparisons extremely difficult. Physically building a sample to test can help make an accurate assessment but is only practicable for a shortlist containing just a few examples. Subsequently making small changes in layout and positioning, to retest and modify the antenna selection and optimise the circuit design, is also time consuming and laborious, particularly if several iterations are needed.

If these issues can be considered at an early stage of the project, designers have a greater opportunity to optimise the system’s RF performance to ensure robust wireless communication and maximise the battery life.

Design automation

Automated design and simulation tools promise to accelerate design evaluations. Conventional EDA tools can provide a representative circuit analysis based on the known behaviour of component models. When it comes to antennas, performance typically depends on factors such as the board layout as well as the effects of surrounding electronic components, any additional antennas nearby, and other materials such as the casing. Many of these factors are unique to a given design, which complicates modelling using conventional tools.

Leveraging machine learning with large data sets, the Oxion interactive framework from Barcelona-based technology innovator Ignion, can now help with antenna selection and design-in to new wireless products. Based on learning from thousands of previous cases, Oxion can help determine the optimum antenna type and PCB dimensions and assess the performance of the chosen antenna in the position described by the user. The tool presents a simple interface for users to enter basic information about the design, and Oxion takes it from there. The tool’s creator, Ignion, already built the Virtual Antenna portfolio comprising six variants which has simplified antenna choices for system designers. Now, by embedding Virtual Antenna IP and additionally integrating APIs to the online portals of major distributors, Oxion can automatically find real antennas whose characteristics and performance correspond closely with the ideal performance as calculated. This can save a large amount of time on product benchmarking and comparison.

Iterate at will

By providing a straightforward and basic description of the design, an engineer using Oxion can quickly generate a proof of concept (PoC) analysis, including expected antenna performance and design recommendations.

Moreover, it is possible to analyse numerous design iterations at will, in real-time, by revising critical antenna parameters such as the position, clearance area and others. Compared to working with a specialist RF-design consultant, taking this type of approach on a human face-to-face basis would be time-consuming and expensive.

Rapid, automated iteration using this tool lets developers explore the impact of various tweaks to the design and in this way determine the optimum solution with respect to all applicable constraints. In addition, its user-friendly interface presents results in an easy-to-understand format, showing the key elements to consider and offering recommendations. The tool also provides the design files including the bill of materials (BoM), including the antenna and matching network parts. This enables the user to take their project forward immediately, as soon as a case is identified that delivers satisfactory results.

Conclusion

The wireless subsystem of an IoT device must usually offer the greatest possible communication range or coverage, for the lowest possible power, all within tight constraints on overall size and space inside the enclosure. Choosing and designing-in the antenna, including determining the optimal position and relative placements of other components, is complex and time consuming. Addressing the issue as early as possible in the project secures the greatest freedom to make any necessary changes quickly and at relatively low cost.

Every project is under time pressure and fast answers are needed. The Oxion platform shows how a new generation of design and simulation tools is emerging, leveraging the latest machine-learning techniques to deliver speed with accuracy. Building on its large database of antennas and design cases, and learning continuously, the tool lets users perform multiple design iterations to find their optimal solution. Its capabilities help non-specialised engineers achieve great RF performance and create certification-ready new products within a short project timeframe.

Oxion is an ever-evolving platform, with a roadmap that prioritises users' needs and is based on AI as a foundational technology. This focus will enable Ignion to continue exploring and implementing cutting-edge AI capabilities that can facilitate the antenna integration process.