v2.1.0

Kalman position estimation using linear Kalman filter with non constant gains

See INS_kf, contributed by @Spyroth in #332.
The filter showed improved performance compared to the complementary filter (Position_estimation). It estimated accelerometer biases, it also estimates altitude (vs origin) and height (vs ground) independently.

Refactoring of Waypoint handler, Mission planner, Navigation and controllers

by @mdouglas90 and @basil-huber in #331
Synchronisation of waypoint list is separated from navigation logic.
Waypoints are treated differenlty based on the attached MAVlink command.
A mission handlers registry was added, it contains a list of mission handlers. Each mission handler can treat specific mission items. For safety, a mission is accept only all its waypoints can be handled.
This architecture also makes it easy to add new behaviour for a new MAVlink command by simply adding a new handler to the handler registry.

v2.0.0

Many changes since last release including:

• Most modules ported to C++
• Hardware abstraction layer (HAL) supporting avr32, stm32 and linux
• Experimental HAL implementation using ChibiOS/HAL backend
• Extended Kalman Filter (EKF) for attitude estimation
• New folder sample_project/ containing example code, always synchronized with the Library
• Continuous integration using travis

v1.5.1

Adds automatic gyro calibration at startup

v1.5.0

Use this version if you want a well-tested version of the library.

This is the last 1.x.x release with all the code written in C. Bug fixes can be back-ported from master by incrementing the PATCH versions (ex: 1.5.1, then 1.5.2, etc...).