openshoe OpenShoe runtime framework changes

- Bugfix of the lossless communication mode which prevented packages from being sent.

John-Olof Nilsson — Tue, 27 Sep 2016 12:29:03 -0000

Bugfix of the lossless communication mode which prevented packages from being sent. - Support for reading magnetometer data from the modules. - Increased speed of I2C communiation. - Time stamping of process sequence. - Support for some experimental hardware and algorithms.
View Changes

Added CAN-bus support. This should not affect the code unless the CAN_INTERFACE macro is defined. A number of ASF components has been added together with the CAN-bus support.

John-Olof Nilsson — Wed, 18 Nov 2015 21:24:04 -0000

Added CAN-bus support. This should not affect the code unless the CAN_INTERFACE macro is defined. A number of ASF components has been added together with the CAN-bus support.
View Changes

Added get_state function.

John-Olof Nilsson — Tue, 10 Nov 2015 18:17:28 -0000

Added get_state function.
View Changes

Bugfix: Return value in case there was no space in the UART buffer was wrong which sometimes caused oscillatory effects when stressing the UART (Bluetooth).

John-Olof Nilsson — Fri, 02 Oct 2015 08:10:22 -0000

Bugfix: Return value in case there was no space in the UART buffer was wrong which sometimes caused oscillatory effects when stressing the UART (Bluetooth).
View Changes

- Upgrade of project files from Atmel Studio 6.5 to 7.

John-Olof Nilsson — Wed, 30 Sep 2015 14:05:15 -0000

Upgrade of project files from Atmel Studio 6.5 to 7. - Upgrade of ASF to 3.18.1 from some old version from 2011. - Clean-up of OpenShoe classic stuff such as SPI routines and external interrupt functionality.
View Changes

Moved and joined communication logic from bluetooth_interface.c and usb_interface.c to external_interface.c. Now the interface resources are bundled in "external interface"-structs and the logic handl...

John-Olof Nilsson — Tue, 29 Sep 2015 14:07:28 -0000

Moved and joined communication logic from bluetooth_interface.c and usb_interface.c to external_interface.c. Now the interface resources are bundled in "external interface"-structs and the logic handled by common functions while the hardware dependencies are hidden in bluetooth_interface.c and usb_interface.c. The interfaces now also have identical package queues.
View Changes

- Separated bias estimation and bias compensation into two different process sequence functions.

John-Olof Nilsson — Thu, 24 Sep 2015 13:14:52 -0000

Separated bias estimation and bias compensation into two different process sequence functions. - Introduced one sample delay of the accelerometer measurements to compensate for the built in delay of the gyro measurements. - Stationarity thresholds as states.
View Changes

-Implemented double buffered (package buffer and transmission buffer) interfaces for both USB and BT.

John-Olof Nilsson — Tue, 03 Mar 2015 16:27:46 -0000

-Implemented double buffered (package buffer and transmission buffer) interfaces for both USB and BT. -USB and BT should now behave the same. Lossless communication is now available for USB. -Proper handling of switching between lossy and lossless communication. -Increased buffer sizes and types to handle larger commands. -Integration of acknowledgement reception and transmission in the normal transmission pipeline. -Various cleanup, small bug fixes and parameter changes. -Removed NACK transmission completely.
View Changes

SIGNIFICANT BUG FIX!!!

John-Olof Nilsson — Mon, 09 Feb 2015 22:13:05 -0000

SIGNIFICANT BUG FIX!!! - The conditional restore process function in the step wise dead reckoning command was set in slot 5 rather than the last position. This meant that a measurement update was run without a any time update which flipped the quarternion. This has now been corrected. - Changed the set_last_process_sequence_element()-function to take a id rather than a pointer. - Removed some TODO-comment which had been done.
View Changes

- Moved IMU_read() to the processing sequence. It's there by default and all commands messing with the process sequence have been complemented to include the IMU_read function. However, if you mess wi...

John-Olof Nilsson — Fri, 06 Feb 2015 14:46:20 -0000

Moved IMU_read() to the processing sequence. It's there by default and all commands messing with the process sequence have been complemented to include the IMU_read function. However, if you mess with the process sequence manually, the IMU_read function need to be handled. - Turned around the "set lossy transmission" logics to "set lossless transmission". No implications for the user interface. - Removed the possibility to set a process sequence slot manually with a pointer. - Made sure that the size limits are respected in all accesses to the state/process function/command info arrays. I've also decreased the limits not to take up unnecessary memory and loops. - Removed command_header_table. The role is fullfilled by command_info_array instead.
View Changes