Boundary-layer glidersonde

The idea of a radiosonde that incorporates abilities of radio-controlled airplanes has been around for a long time. The vision is a radiosonde that is not disposable, but instead uses wings to glide back to the starting point or some other designated landing site to be recovered and reused. The objective is to save costs, especially if more expensive payload is attached, such as particle, ozone or radiation sensors. The concept is naturally called a “glidersonde”.

Prior work

NSSL at NOAA investigated the idea in the 90’s. The AMS published a piece on the concept in their magazine BAMS in 2014 under the title “A Viable Alternative for Conducting Cost Effective Daily Atmospheric Soundings in Developing Countries“.


A prototype of DataBird from GPS Boomerang

In New Zealand, the company GPS Boomerang developed a working version named “DataBird” sometime between 2000 and 2012. Development and sales of the model has however ceased. 

One stumbling block discussed in the literature is the regulatory difficulties of flying anything with self-control, even without active propulsion. Another problem cited is to find a business model that makes a reusable product interesting for companies currently selling a large volume of disposable radiosondes. A third problem is the span of competences needed to both manage aircraft design, sensors and meteorology.

Other projects worth mentioning is CICADA from NRL at the US navy, a prototype for military surveillance, and MAVIS Agnes from University of Southampton, United Kingdom, a prototype of a paper aircraft for atmospheric research.

Windsond on map

Yellow pins along the line of selectable Windsond landing locations, with the tech available today

Relation to Windsond

Windsond already lends some of the philosophy from the fabled glidersonde, given our focus on choosing landing site and recovering the radiosonde. The landing site can be selected along a geographical line defined by the current winds and the sonde is typically recovered within a 20-100 m radius from the predicted landing location.

But what if we could give the user a wide area wherein to select a landing site instead of a line, and what if we could further improve the landing accuracy? That would make the sondes easier to reuse.

Mockup of Windsond miniature glidersonde

Mockup of Windsond miniature glidersonde

Our work

During the spring of 2016, a thesis project at Sparv has investigated the feasibility of giving Windsond an aerodynamic shape and control surfaces, to expand the current line of possible landing locations to a big area. While earlier glidersonde discussions assume a target altitude of 30 000 m, we stick with the Windsond specs of 8 000 m to retain the extremely low weight, portability and ease of use.

The investigation shows we can achieve a big radius of selectable landing sites from the point of cut-down from the balloon. With low winds, the sonde can glide back to the launch site. At medium or strong winds, the sonde naturally needs to land somewhere downwind. It’s technically reasonable to achieve these flight characteristics with a 15 gram glidersonde. The technical analysis shows that all necessary parts fit in the tight space and weight budget. A glidersonde with mechanical actuators is naturally more expensive to manufacture than an ordinary radiosonde, but in reasonable volume the costs are manageable.


We still need to install suitable actuators and electronics and perform necessary tests and certifications. We don’t have a time plan or budget for this work until we find a customer to co-fund it. Economically, the savings possible for a single major field campaign would make this feasible. We hope to realize this concept sooner or later, to further push the limits of cost and convenience for boundary layer data collection.

EDIT June 10: ETH Zurich and Meteoswiss has carried out successful tests with a high-altitude glider radiosonde, using COTS components.

(c) 2015 Sparv Embedded AB
Sparv Embedded AB
Ågatan 39
58222 Linköping, Sweden
Phone: +46 (0)707 312608