Add expected size check for AR00 DL00 command from UAM protocol specification. #106
Conversation
| @@ -102,6 +102,9 @@ struct SerialConnection | |||
| int serial_baud; | |||
| }; | |||
|
|
|||
| static const size_t AR00_PACKET_SIZE = 4379; | |||
| static const size_t DL00_PACKET_SIZE = 1936; | |||
There was a problem hiding this comment.
Are these actually the same for all Hokuyo lasers? That seems unlikely to me, since different lasers have different number of readings per scan
There was a problem hiding this comment.
My understanding is these commands (AR00 and DL00) are separate from the SCIP2.0 protocol which the urg_library is based on and are exclusively part of the UAM protocol specification (for safety LIDARs of which the UAM-05LP is the only one). I've attached the protocol spec.
UAM05LP_CommunicationSpecification_en_2_4_0_r1.pdf
Hence my confidence in setting these as fixed values as anyone using these detailed status commands would need to be doing it with a UAM-05LP scanner as non-safety scanners wouldn't support the UAM commands. It's a bit of a hack the way it's been implemented at the moment effectively switching protocols mid stream to get the safety data and then going back to SCIP to get the scan data again.
Ideally for safety scanners you'd switch to using the UAM protocol exclusively but that's a much bigger project as it would be good to integrate it into the urg_c library.
There was a problem hiding this comment.
Happy to leave it open for people to chime in with other experiences.
As a middle ground we can also set the limit to the maximum string index AR00 and DL00 update functions.
There was a problem hiding this comment.
That seems reasonable - I'm traveling for a few days - but I will test this out on the couple different Hokuyos I have when I get home next week to make sure it works as expected on the lower end lasers
There was a problem hiding this comment.
I've got a UST-10LX at home and ran a quick check, as expected with detailed_status = true the command fails and returns an empty packet. You do get some nasty socket errors though and node doesn't recover well.
I just pushed a quick product name get to the detailed_status branch of the getStatus() function which gives a nice friendly warning if the command isn't supported.
Let me know what you think?
I was a bit slow when I fixed this issue and didn't think to add a check for the message length so there's still a potential for it to fail with a std::length error. This PR adds a packet length check for the AR00 and DL00 from the UAM protocol specification.
The UAM manual doesn't seem to be easily available online so I've added some screen grabs of the two commands for reference:
AR00
DL00