From: Ricardo T. <tel...@gm...> - 2007-08-28 07:00:54
|
Yes,we do care!. Thanks for posting the solution you found. Ricardo 2007/8/28, Deepak Ramachandran <dra...@ui...>: > > All right, I don't know if anyone cares, but I fixed this problem > myself. Basically the AMCL particle filter by default updates the > particles only if the action sensor senses an angular motion of pi/6 > or more, and this is way too much. > You can change this setting by using the update_thresh parameter in > the config file. > > ~Deepak > > > ******************* > > One fish > two fish > Red fish > Blue fish. > > Black fish > Blue fish > Old fish > New fish > > -- Dr Seuss > > > On Aug 27, 2007, at 11:16 AM, Deepak Ramachandran wrote: > > > Hello Everybody, > > > > I posted a message while back asking for help with the > > localizer. Here's a simpler source file that illustrates the same > > problem. All I do is ask the planner to rotate by 180 degrees and > > then query the localizer for the pose and query stage for the true > > angle. I get numbers that differ by up to 40%. Could someone please > > tell what I'm doing wrong? > > > > ~Deepak > > > > #include <stdio.h> > > #include <libplayerc/playerc.h> > > > > playerc_client_t *client; > > playerc_planner_t *planner; > > playerc_localize_t *localizer; > > playerc_position2d_t *position2d; > > > > int initialize() > > { > > // Create a client object and connect to the server; the server > > must > > // be running on "localhost" at port 6665 > > client = playerc_client_create(NULL, "localhost", 6665); > > if (playerc_client_connect(client) != 0) > > { > > fprintf(stderr, "error: %s\n", playerc_error_str()); > > return -1; > > } > > > > // Create a planner proxy (device id "planner:0") and susbscribe > > // in read/write mode > > planner= playerc_planner_create (client, 0); > > > > if (playerc_planner_subscribe(planner, PLAYERC_OPEN_MODE) != 0) > > { > > fprintf(stderr, "error: %s\n", playerc_error_str()); > > return -1; > > } > > > > localizer= playerc_localize_create (client, 0); > > > > if (playerc_localize_subscribe(localizer, PLAYERC_OPEN_MODE) != 0) > > { > > fprintf(stderr, "error: %s\n", playerc_error_str()); > > return -1; > > } > > > > > > position2d = playerc_position2d_create(client, 0); > > if (playerc_position2d_subscribe(position2d, PLAYER_OPEN_MODE)) > > return -1; > > > > return 0; > > } > > > > int main() > > { > > FILE *posfile; > > int started= 0; > > > > if (initialize() != 0) > > return -1; > > > > playerc_planner_set_cmd_pose (planner, 0, 0, 3.14); > > > > while (1) > > { > > playerc_client_read (client); > > > > if ((planner->path_done) || (!started)) > > { > > started= 1; > > > > while (1) > > { > > playerc_client_read(client); > > // printf("in motion: %f\n", localizer->hypoths > > [0].mean.pa); > > if (planner->path_done) > > { > > printf("done: %f %f %f %f\n", > localizer->hypoths[0].mean.px, > > localizer->hypoths[0].mean.py, localizer->hypoths[0].mean.pa, > > position2d->pa); > > exit(0); > > } > > sleep(1); > > } > > > > } > > } > > } > > > > > > > > > > > > ******** > > Starbuck to all Vipers.Do not fire! Repeat, do not fire! > > I'm a friendly OK. We're all friendlies, So let's just, be--friendly! > > > > -- Battlestar Galactica > > > > > > On Aug 27, 2007, at 8:05 AM, don...@gm... wrote: > > > >> Hi, > >> Which player are you using? > >> Thanks. > >> > >> dongqing > >> > >> > >> Deepak Ramachandran wrote: > >>> > >>> Hi, > >>> > >>> I am an intern at the Honda Research Institute and I'm > >>> trying to > >>> use player on pioneer robot. The first thing I'm trying to do is to > >>> get the localizer/planner that comes with player to work > >>> in a simulated environment using stage. There seems to be errors in > >>> the localizer that makes it unusable. I'm not sure if it's a > >>> inherent > >>> weakness in the algorithm or something I'm not doing right. Whenever > >>> I give a command to the planner to move to a specific pose (say > >>> (0,0,0) to (2,1, 0.75) , it typically moves to a position that > >>> undershoots marginally (like (1.8, 0.8, 0.6). This is something I > >>> can > >>> deal with. But the real problem is that when I query the localizer, > >>> the (x,y) part of the pose is pretty accurate, but the angle is > >>> grossly wrong (like half to a whole radian). It doesnt seem like the > >>> localizer has got lost, because the planner still manages to find > >>> its > >>> way and when I look at playernav, the robot icon seems to be > >>> pointing > >>> in the right direction. > >>> > >>> Is there something I'm doing wrong? I've put up some sample code I > >>> wrote (along with all the files needed to get it to run) here: > >>> > >>> http://magma.cs.uiuc.edu/deepak/deepak-example.tar.gz > >>> > >>> It should move the robot to (2,1, 1.57) but when I query the pose > >>> at the end I get something like (1.8,1, 1.57). > >>> > >>> ~Deepak > >>> > >>> > >>> > >>> > >>> ************ > >>> "I believe in the third dimension, but not the other two" > >>> -- Kibo > >>> > >>> > >>> > >>> -------------------------------------------------------------------- > >>> - > >>> ---- > >>> This SF.net email is sponsored by: Splunk Inc. > >>> Still grepping through log files to find problems? Stop. > >>> Now Search log events and configuration files using AJAX and a > >>> browser. > >>> Download your FREE copy of Splunk now >> http://get.splunk.com/ > >>> _______________________________________________ > >>> Playerstage-users mailing list > >>> Pla...@li... > >>> https://lists.sourceforge.net/lists/listinfo/playerstage-users > >>> > >>> > >> Quoted from: > >> http://www.nabble.com/help-with-localizer-tf4251218.html#a12099530 > >> > > > > > > ---------------------------------------------------------------------- > > --- > > This SF.net email is sponsored by: Splunk Inc. > > Still grepping through log files to find problems? Stop. > > Now Search log events and configuration files using AJAX and a > > browser. > > Download your FREE copy of Splunk now >> http://get.splunk.com/ > > _______________________________________________ > > Playerstage-users mailing list > > Pla...@li... > > https://lists.sourceforge.net/lists/listinfo/playerstage-users > > > ------------------------------------------------------------------------- > This SF.net email is sponsored by: Splunk Inc. > Still grepping through log files to find problems? Stop. > Now Search log events and configuration files using AJAX and a browser. > Download your FREE copy of Splunk now >> http://get.splunk.com/ > _______________________________________________ > Playerstage-users mailing list > Pla...@li... > https://lists.sourceforge.net/lists/listinfo/playerstage-users > -- __________________________ Ricardo A. Tellez Research Engineer Pal Robotics phone: +34 93 414 53 47 email: tel...@gm... web: http://www.pal-robotics.com __________________________ |