Mirrors_Framework/g2
1
0
Fork 0
mirror of https://github.com/synthetos/g2.git synced 2026-02-06 11:11:57 +08:00
Code Issues Actions Packages Projects Releases Wiki Activity
Files
c290bdbbcff4eff5aa130447274d79cf8a14f3fb
g2/g2core/planner.h
Rob Giseburt 83e3c7aecb Make sure the primar planner queue is HOT_DATA
2019-10-16 23:32:07 -05:00

674 lines
33 KiB
C++
Raw Blame History

/*
* planner.h - cartesian trajectory planning and motion execution
* This file is part of the g2core project
*
* Copyright (c) 2013 - 2019 Alden S. Hart, Jr.
* Copyright (c) 2013 - 2019 Robert Giseburt
*
* This file ("the software") is free software: you can redistribute it and/or modify
* it under the terms of the GNU General Public License, version 2 as published by the
* Free Software Foundation. You should have received a copy of the GNU General Public
* License, version 2 along with the software. If not, see <http://www.gnu.org/licenses/>.
*
* As a special exception, you may use this file as part of a software library without
* restriction. Specifically, if other files instantiate templates or use macros or
* inline functions from this file, or you compile this file and link it with other
* files to produce an executable, this file does not by itself cause the resulting
* executable to be covered by the GNU General Public License. This exception does not
* however invalidate any other reasons why the executable file might be covered by the
* GNU General Public License.
*
* THE SOFTWARE IS DISTRIBUTED IN THE HOPE THAT IT WILL BE USEFUL, BUT WITHOUT ANY
* WARRANTY OF ANY KIND, EXPRESS OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES
* OF MERCHANTABILITY, FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT
* SHALL THE AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
* LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, OUT OF
* OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE.
*/
/*x
* --- Planner Background ---
*
* The planner is a complicated beast that takes a lot of things into account.
* Planner documentation is scattered about and co-located with the functions
* that perform the actions. Key files are:
*
* - planner.h - This file has defines, structures and prototypes. What you would expect
* - planner.cpp - Core and common functions, queue handling, JSON and command handlers
* - plan_line.cpp - Move planning and queuing, backward planning functions
* - plan_zoid.cpp - Move forward planning, velocity contour calculations and crazy math
* - plan_exec.cpp - Runtime execution functions, calls zoid's forward planning functions
* - stepper.cpp/h - Real-time step generation, segment loading, pulls from plan_exec
* - plan_arc.cpp/h- Arc calculation and runtime functions - layer above the rest of this
*
* --- Planner Overview ---
*
* At high level the planner's job is to reconstruct smooth motion from a set of linear
* approximations while observing and operating within the physical constraints of the
* machine and the physics of motion. Gcode - which consists of a series of into linear
* motion segments - is interpreted, queued to the planner, and joined together to produce
* continuous, synchronized motion. Non-motion commands such as pauses (dwells) and
* peripheral controls such as spindles can also be synchronized in the queue. Arcs are
* just a special case consisting of many linear moves. Arcs are not interpreted directly.
*
* The planner sits in the middle of three system layers:
* - The Gcode interpreter and canonical machine (the 'model'), which feeds...
* - The planner - taking generic commands from the model and queuing them for...
* - The runtime layer - pulling from the planner and driving stepper motors or other devices
*
* The planner queue is the heart of the planner. It's a circular list of ~48 complex structures
* that carry the state of the system needs to execute a linear motion, run a pre-planned command,
* like turning on a spindle, or executing an arbitrary JSON command such as an active comment.
*
* The queue can be viewed as a list of instructions that will execute in exact sequence.
* Some instructions control motion and need to be joined to their forward and backwards
* neighbors so that position, velocity, acceleration, and jerk constraints are not
* violated when moving from one motion to the next.
*
* Others are "commands" that are actually just function callbacks that happen to execute
* at a particular point in time (synchronized with motion commands). Commands can control
* anything you can reasonably program, such as digital IO, serial communications, or
* interpreted commands encoded in JSON.
*
* The buffers in the planner queue are treated as a 'closure' - with all state needed for
* proper execution carried in the planner structure. This is important as it keeps
* model state coherent in a heavily pipelined system. The local copy of the Gcode
* model is carried in the gm structure that is part of each planner buffer.
* See header notes in planner.cpp for more details.
*
* The planner is entered by calling one of:
* - mp_aline() - plan and queue a move with acceleration management
* - mp_dwell() - plan and queue a pause (dwell) to the planner queue
* - mp_queue_command() - queue a canned command
* - mp_json_command() - queue a JSON command for run-time interpretation and execution (M100)
* - mp_json_wait() - queue a JSON wait for run-time interpretation and execution (M101)
* -
* In addition, cm_arc_feed() valaidates and sets up a arc paramewters and calls mp_aline()
* repeatedly to spool out the arc segments into the planner queue.
*
* All the above queueing commands other than mp_aline() are relatively trivial; they just
* post callbacks into the next available planner buffer. Command functions are in 2 parts:
* the part that posts to the queue, and the callback that is executed when the command is
* finally reached in the queue - the _exec().
*
* All mp_aline() does is some preliminary math and then posts an initialized buffer to
* the planner queue. The rest of the move planning operations takes place in background;
* via mp_planner_callback() called from the main loop, and as 'pulls' from the runtime
* stepper operations.
*
* Motion planning is separated into backward planning and forward planning stages.
* Backward planning is initiated by mp_planner_callback() which is called repeatedly
* from the main loop. Backwards planning is performed by mp_plan_block_list() and
* _plan_block(). It starts at the most recently arrived Gcode block. Backward
* planning can occur multiple times for a given buffer, as new moves arriving
* can make the motion profile more optimal.
*
* Backward planning uses velocity and jerk constraints to set maximum entry,
* travel (cruise) and exit velocities for the moves in the queue. In addition,
* it observes the maximum cornering velocities that adjoining moves can sustain
* in a corner or a 'kink' to ensure that the jerk limit of any axis participating
* in the move is not violated. See mp_planner_callback() header comments for more detail.
*
* Forward planning is performed just-in-time and only once, right before the
* planner runtime needs the next buffer. Forward planning provides the final
* contouring of the move. It is invoked by mp_forward_plan() and executed by
* mp_calculate_ramps() in plan_zoid.cpp.
*
* Planner timing operates at a few different levels:
*
* - New lines of ASCII containing commands and moves arriving from the USB are
* parsed and executed as the lowest priority background task from the main loop.
*
* - Backward planning is invoked by a main loop callback, so it also executes as
* a background task, albeit a higher priority one.
*
* - Forward planning and the ultimate preparation of the move for the runtime runs
* as an interrupt as a 'pull' from the planner queue that uses a series of
* interrupts at progressively lower priorities to ensure that the next planner
* buffer is ready before the runtime runs out of forward-planned moves and starves.
*
* Some other functions performed by the planner include:
*
* - Velocity throttling to ensure that very short moves do not execute faster
* than the serial interface can deliver them
*
* - Feed hold and cycle start (resume) operations
*
* - Feed rate override functions and replanning
*
* Some terms that are useful that we try to use consistently:
*
* - buffer - in this context a planner buffer holding a move or a command: mb._ or bf
* - block - a data structure for planning or runtime control. See mp_calculate_ramps() comments
* - move - a linear Gcode move, typically from a G0 or G1 code
* - command - a non-move executable in the planner
* - group - a collection of moves or commands that are treated as a unit
* - line - a line of ASCII gcode or arbitrary text
* - bootstrap - the startup period where the planner collects moves but does not yet execute them
*/
#ifndef PLANNER_H_ONCE
#define PLANNER_H_ONCE
#include "canonical_machine.h" // used for GCodeState_t
#include "hardware.h" // for MIN_SEGMENT_MS
using Motate::Timeout;
/*
* Enums and other type definitions
* All the enums that equal some starting value must be that value. Don't change them
*/
typedef void (*cm_exec_t)(float[], bool[]); // callback to canonical_machine execution function
typedef enum { // planner operating state
PLANNER_IDLE = 0, // planner and movement are idle
PLANNER_STARTUP, // ingesting blocks before movement is started
PLANNER_PRIMING, // preparing new moves for planning ("stitching")
PLANNER_BACK_PLANNING // actively backplanning all blocks, from the newest added to the running block
} plannerState;
typedef enum { // bf->buffer_state values in incresing order so > and < can be used
MP_BUFFER_EMPTY = 0, // buffer is available for use (MUST BE 0)
MP_BUFFER_INITIALIZING, // buffer has been checked out and is being initialzed by aline() or a command
MP_BUFFER_NOT_PLANNED, // planning is in progress - at least vmaxes have been set
MP_BUFFER_BACK_PLANNED, // buffer ready for final planning; velocities have been set
MP_BUFFER_FULLY_PLANNED, // buffer fully planned. May still be replanned
MP_BUFFER_RUNNING, // current running buffer
} bufferState;
typedef enum { // bf->block_type values
BLOCK_TYPE_NULL = 0, // MUST=0 null move - does a no-op
BLOCK_TYPE_ALINE = 1, // MUST=1 acceleration planned line
BLOCK_TYPE_COMMAND = 2, // MUST=2 general command
// All other non-move commands are > BLOCK_TYPE_COMMAND
BLOCK_TYPE_DWELL, // Gcode dwell
BLOCK_TYPE_JSON_WAIT, // JSON wait command
BLOCK_TYPE_TOOL, // T command (T, not M6 tool change)
BLOCK_TYPE_SPINDLE_SPEED, // S command
BLOCK_TYPE_STOP, // program stop
BLOCK_TYPE_END // program end
} blockType;
typedef enum {
BLOCK_INACTIVE = 0, // block is inactive (MUST BE ZERO)
BLOCK_INITIAL_ACTION, // initial value if you need an initialization
BLOCK_ACTIVE // run state
} blockState;
typedef enum {
SECTION_HEAD = 0, // acceleration
SECTION_BODY, // cruise
SECTION_TAIL // deceleration
} moveSection;
#define SECTIONS 3
typedef enum {
SECTION_OFF = 0, // section inactive
SECTION_NEW, // uninitialized section
SECTION_RUNNING // initilized and running
} sectionState;
typedef enum { // code blocks for planning and trapezoid generation
NO_HINT = 0, // block is not hinted
COMMAND_BLOCK, // this block is a command
PERFECT_ACCELERATION, // head-only acceleration at jerk or cannot be improved
PERFECT_DECELERATION, // tail-only deceleration at jerk or cannot be improved
PERFECT_CRUISE, // body-only cruise: Ve = Vc = Vx != 0
MIXED_ACCELERATION, // kinked acceleration reaches and holds cruise (HB)
MIXED_DECELERATION, // kinked deceleration starts with a cruise region (BT)
// The rest are for reporting from the zoid function what was selected.
ZERO_VELOCITY, // Ve = Vc = Vx = 0
ZERO_BUMP, // Ve = Vx = 0, Vc != 0
SYMMETRIC_BUMP, // (Ve = Vx) < Vc
ASYMMETRIC_BUMP, // (Ve != Vx) < Vc
} blockHint;
/*** Most of these factors are the result of a lot of tweaking. Change with caution.***/
#ifdef PLANNER_BUFFER_POOL_SIZE
#error Please change PLANNER_BUFFER_POOL_SIZE define to PLANNER_QUEUE_SIZE (found elsewhere, unfortunately)
#endif
#ifndef PLANNER_QUEUE_SIZE
#define PLANNER_QUEUE_SIZE ((uint8_t)48) // Suggest 12 min. Limit is 255
#endif
#ifndef SECONDARY_QUEUE_SIZE
#define SECONDARY_QUEUE_SIZE ((uint8_t)12) // Secondary planner queue for feedhold operations
#endif
#define PLANNER_BUFFER_HEADROOM ((uint8_t)4) // Buffers to reserve in planner before processing new input line
#define JERK_MULTIPLIER ((float)1000000) // DO NOT CHANGE - must always be 1 million
#define JUNCTION_INTEGRATION_MIN (0.05) // JT minimum allowable setting
#define JUNCTION_INTEGRATION_MAX (5.00) // JT maximum allowable setting
#ifndef MIN_SEGMENT_MS // boards can override this value in hardware.h
#define MIN_SEGMENT_MS ((float)0.75) // minimum segment milliseconds
#endif
#define NOM_SEGMENT_MS ((float)MIN_SEGMENT_MS*2.0) // nominal segment ms (at LEAST MIN_SEGMENT_MS * 2)
#define MIN_BLOCK_MS ((float)MIN_SEGMENT_MS*2.0) // minimum block (whole move) milliseconds
#define BLOCK_TIMEOUT_MS ((float)30.0) // MS before deciding there are no new blocks arriving
#define PHAT_CITY_MS ((float)100.0) // if you have at least this much time in the planner
#define NOM_SEGMENT_TIME ((float)(NOM_SEGMENT_MS / 60000)) // DO NOT CHANGE - time in minutes
#define NOM_SEGMENT_USEC ((float)(NOM_SEGMENT_MS * 1000)) // DO NOT CHANGE - time in microseconds
#define MIN_SEGMENT_TIME ((float)(MIN_SEGMENT_MS / 60000)) // DO NOT CHANGE - time in minutes
#define MIN_BLOCK_TIME ((float)(MIN_BLOCK_MS / 60000)) // DO NOT CHANGE - time in minutes
#define PHAT_CITY_TIME ((float)(PHAT_CITY_MS / 60000)) // DO NOT CHANGE - time in minutes
#define FEED_OVERRIDE_ENABLE false // initial value
#define FEED_OVERRIDE_MIN (0.05) // 5% minimum
#define FEED_OVERRIDE_MAX (2.00) // 200% maximum
#define FEED_OVERRIDE_RAMP_TIME (0.500/60) // ramp time for feed overrides
#define FEED_OVERRIDE_FACTOR (1.00) // initial value
#define TRAVERSE_OVERRIDE_ENABLE false // initial value
#define TRAVERSE_OVERRIDE_MIN (0.05) // 5% minimum
#define TRAVERSE_OVERRIDE_MAX (1.00) // 100% maximum
#define TRAVERSE_OVERRIDE_FACTOR (1.00) // initial value
//// Specialized equalities for comparing velocities with tolerances
//// These determine allowable velocity discontinuities between blocks (among other tests)
//// RG: Simulation shows +-0.001 is about as much as we should allow.
// VELOCITY_EQ(v0,v1) reads: "True if v0 is within 0.0001 of v1"
// VELOCITY_LT(v0,v1) reads: "True if v0 is less than v1 by at least 0.0001"
#define VELOCITY_EQ(v0,v1) ( std::abs(v0-v1) < 0.0001 )
#define VELOCITY_LT(v0,v1) ( (v1 - v0) > 0.0001 )
#define Vthr2 300.0
#define Veq2_hi 10.0
#define Veq2_lo 1.0
#define VELOCITY_ROUGHLY_EQ(v0,v1) ( (v0 > Vthr2) ? std::abs(v0-v1) < Veq2_hi : std::abs(v0-v1) < Veq2_lo )
/* Planner Diagnostics */
//#define __PLANNER_DIAGNOSTICS // comment this out to drop diagnostics
#ifdef __PLANNER_DIAGNOSTICS
#define ASCII_ART(s) xio_writeline(s)
#define UPDATE_BF_DIAGNOSTICS(bf) { bf->linenum = bf->gm.linenum; \
bf->block_time_ms = bf->block_time*60000; \
bf->plannable_time_ms = bf->plannable_time*60000; }
#define UPDATE_MP_DIAGNOSTICS { mp->plannable_time_ms = mp->plannable_time*60000; }
#define SET_PLANNER_ITERATIONS(i) { bf->iterations = i; }
#define INC_PLANNER_ITERATIONS { bf->iterations++; }
#define SET_MEET_ITERATIONS(i) { bf->meet_iterations = i; }
#define INC_MEET_ITERATIONS { bf->meet_iterations++; }
#else
#define ASCII_ART(s)
#define UPDATE_BF_DIAGNOSTICS
#define UPDATE_MP_DIAGNOSTICS
#define SET_PLANNER_ITERATIONS(i)
#define INC_PLANNER_ITERATIONS
#define SET_MEET_ITERATIONS(i)
#define INC_MEET_ITERATIONS
#endif
/*
* Planner structures
*
* Be aware of the distinction between 'buffers' and 'blocks'
* Please refer to header comments in for important details on buffers and blocks
* - plan_zoid.cpp / mp_calculate_ramps()
* - plan_exec.cpp / mp_exec_aline()
*/
//**** Planner Queue Structures ****
typedef struct mpBuffer {
// *** CAUTION *** These two pointers are not reset by _clear_buffer()
struct mpBuffer *pv; // static pointer to previous buffer
struct mpBuffer *nx; // static pointer to next buffer
uint8_t buffer_number; // DIAGNOSTIC for easier debugging
stat_t (*bf_func)(struct mpBuffer *bf); // callback to buffer exec function
cm_exec_t cm_func; // callback to canonical machine execution function
#ifdef __PLANNER_DIAGNOSTICS
uint32_t linenum; // mirror of bf->gm.linenum
int iterations;
float block_time_ms;
float plannable_time_ms; // time in planner
float plannable_length; // length in planner
uint8_t meet_iterations; // iterations needed in _get_meet_velocity
#endif
bufferState buffer_state; // used to manage queuing/dequeuing
blockType block_type; // used to dispatch to run routine
blockState block_state; // move state machine sequence
blockHint hint; // hint the block for zoid and other planning operations. Must be accurate or NO_HINT
// block parameters
float unit[AXES]; // unit vector for axis scaling & planning
bool axis_flags[AXES]; // set true for axes participating in the move & for command parameters
float junction_unit[AXES]; // unit vector delta at the junction for cornering. Needed for groups of small moves.
float junction_length_since; // length total of the moves since the junction_unit was captured. See _calculate_junction_vmax() comments.
bool plannable; // set true when this block can be used for planning
float length; // total length of line or helix in mm
float block_time; // computed move time for entire block (move)
float override_factor; // feed rate or rapid override factor for this block ("override" is a reserved word)
// *** SEE NOTES ON THESE VARIABLES, in aline() ***
// We removed all entry_* values.
// To get the entry_* values, look at pv->exit_* or mr->exit_*
float cruise_velocity; // cruise velocity requested & achieved
float exit_velocity; // exit velocity requested for the move
// is also the entry velocity of the *next* move
float cruise_vset; // cruise velocity requested for move - prior to overrides
float cruise_vmax; // cruise max velocity adjusted for overrides
float exit_vmax; // max exit velocity possible for this move
// is also the maximum entry velocity of the next move
float absolute_vmax; // fastest this block can move w/o exceeding constraints
float junction_vmax; // maximum the exit velocity can be to go through the junction
// between the NEXT BLOCK AND THIS ONE
float jerk; // maximum linear jerk term for this move
float jerk_sq; // Jm^2 is used for planning (computed and cached)
float recip_jerk; // 1/Jm used for planning (computed and cached)
float sqrt_j; // sqrt(jM) used for planning (computed and cached)
float q_recip_2_sqrt_j; // (q/(2 sqrt(jM))) where q = (sqrt(10)/(3^(1/4))), used in length computations (computed and cached)
GCodeState_t gm; // Gcode model state - passed from model, used by planner and runtime
// clears the above structure
void reset() {
bf_func = nullptr;
cm_func = nullptr;
#ifdef __PLANNER_DIAGNOSTICS
linenum = 0;
iterations = 0;
block_time_ms = 0;
plannable_time_ms = 0;
plannable_length = 0;
meet_iterations = 0;
#endif
buffer_state = MP_BUFFER_EMPTY;
block_type = BLOCK_TYPE_NULL;
block_state = BLOCK_INACTIVE;
hint = NO_HINT;
for (uint8_t i = 0; i< AXES; i++) {
unit[i] = 0;
junction_unit[i] = 0;
junction_length_since = 0;
axis_flags[i] = 0;
}
plannable = false;
length = 0.0;
block_time = 0.0;
override_factor = 0.0;
cruise_velocity = 0.0;
exit_velocity = 0.0;
cruise_vset = 0.0;
cruise_vmax = 0.0;
exit_vmax = 0.0;
absolute_vmax = 0.0;
junction_vmax = 0.0;
jerk = 0.0;
jerk_sq = 0.0;
recip_jerk = 0.0;
sqrt_j = 0.0;
q_recip_2_sqrt_j = 0.0;
gm.reset();
}
} mpBuf_t;
typedef struct mpPlannerQueue { // control structure for queue
magic_t magic_start; // magic number to test memory integrity
mpBuf_t *r; // run buffer pointer
mpBuf_t *w; // write buffer pointer
uint8_t queue_size; // total number of buffers, one-based (e.g. 48 not 47)
uint8_t buffers_available; // running count of available buffers in queue
mpBuf_t *bf; // pointer to buffer pool (storage array)
magic_t magic_end;
} mpPlannerQueue_t;
//**** Planner Runtime structures ****
typedef struct mpBlockRuntimeBuf { // Data structure for just the parts of RunTime that we need to plan a BLOCK
struct mpBlockRuntimeBuf *nx; // singly-linked-list
float head_length; // copies of bf variables of same name
float body_length;
float tail_length;
float head_time; // copies of bf variables of same name
float body_time;
float tail_time;
float cruise_velocity; // velocity at the end of the head and the beginning of the tail
float exit_velocity; // velocity at the end of the move
} mpBlockRuntimeBuf_t;
typedef struct mpPlannerRuntime { // persistent runtime variables
// uint8_t (*run_move)(struct mpMoveRuntimeSingleton *m); // currently running move - left in for reference
magic_t magic_start; // magic number to test memory integrity
blockState block_state; // state of the overall move
moveSection section; // what section is the move in?
sectionState section_state; // state within a move section
bool out_of_band_dwell_flag; // set true to conditionally execute out-of-band dwell
float out_of_band_dwell_seconds; // time for out-of-band dwell
float unit[AXES]; // unit vector for axis scaling & planning
bool axis_flags[AXES]; // set true for axes participating in the move
float target[AXES]; // final target for bf (used to correct rounding errors)
float position[AXES]; // current move position
float waypoint[SECTIONS][AXES]; // head/body/tail endpoints for correction
float target_steps[MOTORS]; // current MR target (absolute target as steps)
float position_steps[MOTORS]; // current MR position (target from previous segment)
float commanded_steps[MOTORS]; // will align with next encoder sample (target from 2nd previous segment)
float encoder_steps[MOTORS]; // encoder position in steps - ideally the same as commanded_steps
float following_error[MOTORS]; // difference between encoder_steps and commanded steps
mpBlockRuntimeBuf_t *r; // block that is running
mpBlockRuntimeBuf_t *p; // block that is being planned, p might == r
mpBlockRuntimeBuf_t block[2]; // buffer holding the two blocks
mpBuf_t *plan_bf; // DIAGNOSTIC - pointer to next buffer to plan
mpBuf_t *run_bf; // DIAGNOSTIC - pointer to currently running buffer
float entry_velocity; // entry values for the currently running block
float segments; // number of segments in line (also used by arc generation)
uint32_t segment_count; // count of running segments
float segment_velocity; // computed start velocity for aline segment
float target_velocity; // computed end velocity for aline segment
float segment_time; // actual time increment per aline segment
float forward_diff_1; // forward difference level 1
float forward_diff_2; // forward difference level 2
float forward_diff_3; // forward difference level 3
float forward_diff_4; // forward difference level 4
float forward_diff_5; // forward difference level 5
GCodeState_t gm; // gcode model state currently executing
magic_t magic_end;
// resets mpPlannerRuntime structure without actually wiping it
void reset() {
block_state = BLOCK_INACTIVE;
section = SECTION_HEAD;
section_state = SECTION_OFF;
entry_velocity = 0; // needed to ensure next block in forward planning starts from 0 velocity
r->exit_velocity = 0; // ditto
segment_velocity = 0;
}
} mpPlannerRuntime_t;
//**** Master Planner Structure ***
typedef struct mpPlanner { // common variables for a planner context
magic_t magic_start; // magic number to test memory integrity
// DIAGNOSTICS
float run_time_remaining_ms;
float plannable_time_ms;
// planner position
float position[AXES]; // final move position for planning purposes
// timing variables
float run_time_remaining; // time left in runtime (including running block)
float plannable_time; // time in planner that can actually be planned
// planner state variables
plannerState planner_state; // current state of planner
bool request_planning; // set true to request backplanning
bool backplanning; // true if planner is in a back-planning pass
bool mfo_active; // true if mfo override is in effect
bool ramp_active; // true when a ramp is occurring
bool entry_changed; // mark if exit_velocity changed to invalidate next block's hint
// feed overrides and ramp variables (these extend the variables in cm->gmx)
float mfo_factor; // runtime override factor
float ramp_target;
float ramp_dvdt;
// objects
Timeout block_timeout; // Timeout object for block planning
// planner pointers
mpBuf_t *p; // planner buffer pointer
mpBuf_t *c; // pointer to buffer immediately following critical region
mpBuf_t *planning_return; // buffer to return to once back-planning is complete
mpPlannerRuntime_t *mr; // bind to mr associated with this planner
mpPlannerQueue_t q; // embed a planner buffer queue manager
magic_t magic_end;
// clears mpPlanner structure but leaves position alone
void reset() {
run_time_remaining = 0;
plannable_time = 0;
planner_state = PLANNER_IDLE;
request_planning = false;
backplanning = false;
mfo_active = false;
ramp_active = false;
entry_changed = false;
block_timeout.clear();
}
} mpPlanner_t;
// Reference global scope structures
extern mpPlanner_t *mp HOT_DATA; // currently active planner (global variable)
extern mpPlanner_t mp1 HOT_DATA; // primary planning context
extern mpPlanner_t mp2 HOT_DATA; // secondary planning context
extern mpPlannerRuntime_t *mr HOT_DATA; // context for block runtime
extern mpPlannerRuntime_t mr1 HOT_DATA; // primary planner runtime context
extern mpPlannerRuntime_t mr2 HOT_DATA; // secondary planner runtime context
extern mpBuf_t mp1_queue[PLANNER_QUEUE_SIZE] HOT_DATA; // storage allocation for primary planner queue buffers
extern mpBuf_t mp2_queue[SECONDARY_QUEUE_SIZE]; // storage allocation for secondary planner queue buffers
/*
* Global Scope Functions
*/
//**** planner.cpp functions
void planner_init(mpPlanner_t *_mp, mpPlannerRuntime_t *_mr, mpBuf_t *queue, uint8_t queue_size);
void planner_reset(mpPlanner_t *_mp);
stat_t planner_assert(const mpPlanner_t *_mp);
void mp_halt_runtime(void);
void mp_set_planner_position(uint8_t axis, const float position);
void mp_set_runtime_position(uint8_t axis, const float position);
void mp_set_steps_to_runtime_position(void);
stat_t mp_set_target_steps(const float target_steps[]) HOT_FUNC;
stat_t mp_set_target_steps(const float target_steps[MOTORS], const float start_velocities[MOTORS], const float end_velocities[MOTORS], const float segment_time) HOT_FUNC;
void mp_queue_command(void(*cm_exec)(float *, bool *), float *value, bool *flag);
stat_t mp_runtime_command(mpBuf_t *bf);
stat_t mp_json_command(char *json_string);
stat_t mp_json_command_immediate(char *json_string);
stat_t mp_json_wait(char *json_string);
stat_t mp_dwell(const float seconds);
void mp_end_dwell(void);
void mp_request_out_of_band_dwell(float seconds);
//**** planner functions and helpers
uint8_t mp_get_planner_buffers(const mpPlanner_t *_mp);
bool mp_planner_is_full(const mpPlanner_t *_mp);
bool mp_has_runnable_buffer(const mpPlanner_t *_mp);
bool mp_is_phat_city_time(void);
stat_t mp_planner_callback();
void mp_replan_queue(mpBuf_t *bf);
void mp_start_feed_override(const float ramp_time, const float override);
void mp_end_feed_override(const float ramp_time);
void mp_start_traverse_override(const float ramp_time, const float override);
void mp_end_traverse_override(const float ramp_time);
void mp_planner_time_accounting(void);
//**** planner buffer primitives
//void mp_init_planner_buffers(void);
//mpBuf_t * mp_get_w(int8_t q);
//mpBuf_t * mp_get_r(int8_t q);
mpBuf_t * mp_get_w(void);
mpBuf_t * mp_get_r(void);
//mpBuf_t * mp_get_prev_buffer(const mpBuf_t *bf); // Use the following macro instead
//mpBuf_t * mp_get_next_buffer(const mpBuf_t *bf); // Use the following macro instead
#define mp_get_prev_buffer(b) ((mpBuf_t *)(b->pv))
#define mp_get_next_buffer(b) ((mpBuf_t *)(b->nx))
mpBuf_t * mp_get_write_buffer(void) HOT_FUNC;
void mp_commit_write_buffer(const blockType block_type) HOT_FUNC;
mpBuf_t * mp_get_run_buffer(void) HOT_FUNC;
bool mp_free_run_buffer(void) HOT_FUNC;
//**** plan_line.c functions
void mp_zero_segment_velocity(void); // getters and setters...
float mp_get_runtime_velocity(void);
float mp_get_runtime_absolute_position(mpPlannerRuntime_t *_mr, uint8_t axis);
float mp_get_runtime_display_position(uint8_t axis);
void mp_set_runtime_display_offset(float offset[]);
bool mp_get_runtime_busy(void);
bool mp_runtime_is_idle(void);
stat_t mp_aline(GCodeState_t *_gm); // line planning...
void mp_plan_block_list(void);
void mp_plan_block_forward(mpBuf_t *bf);
//**** plan_zoid.c functions
stat_t mp_calculate_ramps(mpBlockRuntimeBuf_t *block, mpBuf_t *bf, const float entry_velocity);
float mp_get_target_length(const float v_0, const float v_1, const mpBuf_t *bf);
float mp_get_target_velocity(const float v_0, const float L, const mpBuf_t *bf); // acceleration ONLY
float mp_get_decel_velocity(const float v_0, const float L, const mpBuf_t *bf); // deceleration ONLY
float mp_find_t(const float v_0, const float v_1, const float L, const float totalL, const float initial_t, const float T);
float mp_calc_v(const float t, const float v_0, const float v_1); // compute the velocity along the curve accelerating from v_0 to v_1, at position t=[0,1]
float mp_calc_a(const float t, const float v_0, const float v_1, const float T); // compute acceleration over curve accelerating from v_0 to v_1, at position t=[0,1], total time T
float mp_calc_j(const float t, const float v_0, const float v_1, const float T); // compute jerk over curve accelerating from v_0 to v_1, at position t=[0,1], total time T
//float mp_calc_l(const float t, const float v_0, const float v_1, const float T); // compute length over curve accelerating from v_0 to v_1, at position t=[0,1], total time T
//**** plan_exec.c functions
stat_t mp_forward_plan(void);
stat_t mp_exec_move(void);
stat_t mp_exec_aline(mpBuf_t *bf);
void mp_exit_hold_state(void);
void mp_dump_planner(mpBuf_t *bf_start);
#endif // End of include Guard: PLANNER_H_ONCE
Reference in New Issue View Git Blame Copy Permalink