EmbodiedAI
/
RoboWaiter
mirror of https://github.com/HPCL-EI/RoboWaiter.git
3
0
Fork 0
Code Issues Packages Projects Releases Wiki Activity

Merge remote-tracking branch 'origin/main'

This commit is contained in:
wuziji 2023-11-16 20:07:17 +08:00
parent 439e0dce68 8d40d43c5a
commit 38f235eb09
18 changed files with 230 additions and 32 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

2
robowaiter/behavior_lib/_base/Behavior.py
View File

@ -33,6 +33,8 @@ class Bahavior(ptree.behaviour.Behaviour):
'Water': 'Glass',
'Dessert':'Plate'
}
@classmethod
def get_ins_name(cls,*args):

2
robowaiter/behavior_lib/act/Make.py
View File

@ -57,6 +57,6 @@ class Make(Act):
self.scene.state["condition_set"] |= (self.info["add"])
self.scene.state["condition_set"] -= self.info["del_set"]
print("condition_set:",self.scene.state["condition_set"])
# print("condition_set:",self.scene.state["condition_set"])
return Status.RUNNING

6
robowaiter/behavior_lib/act/MoveTo.py
View File

@ -21,7 +21,7 @@ class MoveTo(Act):
info['pre'] |= {f'Exist({arg})'}
info["add"] = {f'At(Robot,{arg})'}
info["del_set"] = {f'At(Robot,{place})' for place in cls.valid_args if place != arg}
info['cost']=5
info['cost']=10
return info
@ -53,8 +53,6 @@ class MoveTo(Act):
if obj.name == target_name:
obj_info = obj_dict[id]
dis = self.scene.cal_distance_to_robot(obj_info.location.X, obj_info.location.Y, obj_info.location.Z)
if id==275:
print("275'dis:",dis)
if dis<min_dis:
min_dis = dis
obj_id = id
@ -64,7 +62,7 @@ class MoveTo(Act):
if obj_id == -1:
return ptree.common.Status.FAILURE
print("self.target_place:",self.target_place,"id:",obj_id,"dis:",min_dis)
# print("self.target_place:",self.target_place,"id:",obj_id,"dis:",min_dis)
self.scene.move_to_obj(obj_id=obj_id)
# 为了演示，写死咖啡位置

4
robowaiter/behavior_lib/act/PutDown.py
View File

@ -22,6 +22,8 @@ class PutDown(Act):
info["pre"] = {f'Holding({arg[0]})',f'At(Robot,{arg[1]})'}
info["add"] = {f'Holding(Nothing)',f'On({arg[0]},{arg[1]})'}
info["del_set"] = {f'Holding({arg[0]})'}
info['cost'] = 100
return info
@ -36,4 +38,6 @@ class PutDown(Act):
self.scene.state["condition_set"] |= (self.info["add"])
self.scene.state["condition_set"] -= self.info["del_set"]
print("After PutDown condition_set:",self.scene.state["condition_set"])
return Status.RUNNING

3
robowaiter/behavior_lib/cond/On.py
View File

@ -16,6 +16,9 @@ class On(Cond):
def _update(self) -> ptree.common.Status:
# if self.scene.status?
# print("self.name:",self.name)
# print("On: condition_set:",self.scene.state["condition_set"])
if self.name in self.scene.state["condition_set"]:
return ptree.common.Status.SUCCESS
else:

30
robowaiter/behavior_tree/obtea/OptimalBTExpansionAlgorithm.py
View File

@ -105,8 +105,8 @@ class OptBTExpAlgorithm:
[copy.deepcopy(pair_node.cond_leaf), copy.deepcopy(pair_node.act_leaf)])
subtree.add_child([copy.deepcopy(sequence_structure)]) # subtree 是回不断变化的，它的父亲是self.bt
# 增加实时条件判断，满足条件就不再扩展
if c <= self.scene.state["condition_set"]:
return True
# if c <= self.scene.state["condition_set"]:
# return True
else:
subtree.add_child([copy.deepcopy(pair_node.act_leaf)])
@ -138,14 +138,14 @@ class OptBTExpAlgorithm:
break
if valid:
# 把符合条件的动作节点都放到列表里
if self.verbose:
print("———— -- %s 符合条件放入列表" % actions[i].name)
c_attr_node = Leaf(type='cond', content=c_attr, mincost=current_mincost + actions[i].cost)
a_attr_node = Leaf(type='act', content=actions[i], mincost=current_mincost + actions[i].cost)
cond_anc_pair = CondActPair(cond_leaf=c_attr_node, act_leaf=a_attr_node)
self.nodes.append(copy.deepcopy(cond_anc_pair)) # condition node list
self.traversed.append(c_attr) # 重点 the set of expanded conditions
# 把符合条件的动作节点都放到列表里
if self.verbose:
print("———— -- %s 符合条件放入列表,对应的c为 %s" % (actions[i].name,c_attr))
if self.verbose:
print("算法结束！\n")
@ -188,13 +188,23 @@ class OptBTExpAlgorithm:
# 树的dfs
def dfs_ptml(self,parnode):
def dfs_ptml(self,parnode,is_root=False):
for child in parnode.children:
if isinstance(child, Leaf):
if child.type == 'cond':
self.ptml_string += "cond "
c_set_str = '\n cond '.join(map(str, child.content)) + "\n"
self.ptml_string += c_set_str
if is_root and len(child.content) > 1:
# 把多个 cond 串起来
self.ptml_string += "sequence{\n"
self.ptml_string += "cond "
c_set_str = '\n cond '.join(map(str, child.content)) + "\n"
self.ptml_string += c_set_str
self.ptml_string += '}\n'
else:
self.ptml_string += "cond "
c_set_str = '\n cond '.join(map(str, child.content)) + "\n"
self.ptml_string += c_set_str
elif child.type == 'act':
if '(' not in child.content.name:
self.ptml_string += 'act ' + child.content.name + "()\n"
@ -212,7 +222,7 @@ class OptBTExpAlgorithm:
def get_ptml(self):
self.ptml_string = "selector{\n"
self.dfs_ptml(self.bt.children[0])
self.dfs_ptml(self.bt.children[0],is_root=True)
self.ptml_string += '}\n'
return self.ptml_string

2
robowaiter/behavior_tree/obtea/opt_bt_exp_main.py
View File

@ -31,7 +31,7 @@ class BTOptExpInterface:
:return: A PTML string representing the outcome of the behavior tree.
"""
self.goal = goal
self.algo = OptBTExpAlgorithm(verbose=True)
self.algo = OptBTExpAlgorithm(verbose=False)
self.algo.clear()
self.algo.run_algorithm(self.goal, self.actions,self.scene) # 调用算法得到行为树保存至 algo.bt
self.ptml_string = self.algo.get_ptml()

2
robowaiter/llm_client/data/test_questions.txt
View File

@ -1 +1 @@
{"测试VLM：做一杯咖啡": {"Answer": "测试VLM：做一杯咖啡", "Goal": "{\"On(Coffee,CoffeeTable)\"}"}, "测试VLM：做一杯咖啡放到吧台上": {"Answer": "测试VLM：做一杯咖啡放到吧台上", "Goal": "{\"On(Coffee,Bar)\"}"}, "测试VLM：做一杯咖啡放到水杯桌上并倒水": {"Answer": "测试VLM：做一杯咖啡放到水杯桌上并倒水", "Goal": "{\"On(Coffee,WaterTable)\"}"}, "测试VLN：前往2号桌": {"Answer": "测试VLN：前往2号桌", "Goal": "{\"At(Robot,Table2)\"}"}, "测试AEM": {"Answer": "测试AEM", "Goal": "{\"EnvExplored()\"}"}, "测试VLM：倒一杯水": {"Answer": "测试VLM：倒一杯水", "Goal": "{\"On(Water,WaterTable)\"}"}, "测试VLM：开空调": {"Answer": "测试VLM：开空调", "Goal": "{\"Is(AC,On)\"}"}, "测试VLM：关空调": {"Answer": "测试VLM：关空调", "Goal": "{\"Is(AC,Off)\"}"}, "测试VLM：关大厅灯": {"Answer": "测试VLM：关大厅灯", "Goal": "{\"Is(HallLight,Off)\"}"}, "测试VLM：开大厅灯": {"Answer": "测试VLM：开大厅灯", "Goal": "{\"Is(HallLight,On)\"}"}, "测试VLM：关筒灯": {"Answer": "测试VLM：关筒灯", "Goal": "{\"Is(TubeLight,Off)\"}"}, "测试VLM：开筒灯": {"Answer": "测试VLM：开筒灯", "Goal": "{\"Is(TubeLight,On)\"}"}, "测试VLM：关窗帘": {"Answer": "测试VLM：关窗帘", "Goal": "{\"Is(Curtain,Off)\"}"}, "测试VLM：开窗帘": {"Answer": "测试VLM：开窗帘", "Goal": "{\"Is(Curtain,On)\"}"}, "测试VLM：拖地": {"Answer": "测试VLM：拖地", "Goal": "{\"Is(Floor,Clean)\"}"}, "测试VLM：擦桌子": {"Answer": "测试VLM：擦桌子", "Goal": "{\"Is(Table1,Clean)\"}"}, "测试VLM：整理椅子": {"Answer": "测试VLM：整理椅子", "Goal": "{\"Is(Chairs,Clean)\"}"}, "测试VLM：把冰红茶放到Table2": {"Answer": "测试VLM：把冰红茶放到Table2", "Goal": "{\"On(BottledDrink,Table2)\"}"}, "我有点热，能开个空调吗？": {"Answer": "当然可以，我现在就开！", "Goal": "{\"Is(AC,On)\"}"}, "可以带我去吗": {"Answer": "当然可以，前往一号桌", "Goal": "{\"At(Robot,Table1)\"}"}}
{"测试VLM：做一杯咖啡": {"Answer": "测试VLM：做一杯咖啡", "Goal": "{\"On(Coffee,CoffeeTable)\"}"}, "测试VLM：做一杯咖啡放到吧台上": {"Answer": "测试VLM：做一杯咖啡放到吧台上", "Goal": "{\"On(Coffee,Bar)\"}"}, "测试VLM：做一杯咖啡放到水杯桌上，再倒一杯水": {"Answer": "测试VLM：做一杯咖啡放到水杯桌上，再倒一杯水", "Goal": "{\"On(Coffee,WaterTable)\",\"On(Water,WaterTable)\"}"}, "测试VLN：前往2号桌": {"Answer": "测试VLN：前往2号桌", "Goal": "{\"At(Robot,Table2)\"}"}, "测试AEM": {"Answer": "测试AEM", "Goal": "{\"EnvExplored()\"}"}, "测试VLM：倒一杯水": {"Answer": "测试VLM：倒一杯水", "Goal": "{\"On(Water,WaterTable)\"}"}, "测试VLM：开空调": {"Answer": "测试VLM：开空调", "Goal": "{\"Is(AC,On)\"}"}, "测试VLM：关空调": {"Answer": "测试VLM：关空调", "Goal": "{\"Is(AC,Off)\"}"}, "测试VLM：关大厅灯": {"Answer": "测试VLM：关大厅灯", "Goal": "{\"Is(HallLight,Off)\"}"}, "测试VLM：开大厅灯": {"Answer": "测试VLM：开大厅灯", "Goal": "{\"Is(HallLight,On)\"}"}, "测试VLM：关筒灯": {"Answer": "测试VLM：关筒灯", "Goal": "{\"Is(TubeLight,Off)\"}"}, "测试VLM：开筒灯": {"Answer": "测试VLM：开筒灯", "Goal": "{\"Is(TubeLight,On)\"}"}, "测试VLM：关窗帘": {"Answer": "测试VLM：关窗帘", "Goal": "{\"Is(Curtain,Off)\"}"}, "测试VLM：开窗帘": {"Answer": "测试VLM：开窗帘", "Goal": "{\"Is(Curtain,On)\"}"}, "测试VLM：拖地": {"Answer": "测试VLM：拖地", "Goal": "{\"Is(Floor,Clean)\"}"}, "测试VLM：擦桌子": {"Answer": "测试VLM：擦桌子", "Goal": "{\"Is(Table1,Clean)\"}"}, "测试VLM：整理椅子": {"Answer": "测试VLM：整理椅子", "Goal": "{\"Is(Chairs,Clean)\"}"}, "测试VLM：把冰红茶放到Table2": {"Answer": "测试VLM：把冰红茶放到Table2", "Goal": "{\"On(BottledDrink,Table2)\"}"}, "我有点热，能开个空调吗？": {"Answer": "当然可以，我现在就开！", "Goal": "{\"Is(AC,On)\"}"}, "可以带我去吗": {"Answer": "当然可以，前往一号桌", "Goal": "{\"At(Robot,Table1)\"}"}}

2
robowaiter/llm_client/data_raw/test_questions.csv
View File

@ -1,7 +1,7 @@
Question,Answer,Goal
测试VLM：做一杯咖啡,测试VLM：做一杯咖啡,"{""On(Coffee,CoffeeTable)""}"
测试VLM：做一杯咖啡放到吧台上,测试VLM：做一杯咖啡放到吧台上,"{""On(Coffee,Bar)""}"
测试VLM：做一杯咖啡放到水杯桌上并倒水,测试VLM：做一杯咖啡放到水杯桌上并倒水,"{""On(Coffee,WaterTable)""}"
测试VLM：做一杯咖啡放到水杯桌上，再倒一杯水,测试VLM：做一杯咖啡放到水杯桌上，再倒一杯水,"{""On(Coffee,WaterTable)"",""On(Water,WaterTable)""}"
测试VLN：前往2号桌,测试VLN：前往2号桌,"{""At(Robot,Table2)""}"
测试AEM,测试AEM,"{""EnvExplored()""}"
测试VLM：倒一杯水,测试VLM：倒一杯水,"{""On(Water,WaterTable)""}"

1 Question Answer Goal
2 测试VLM：做一杯咖啡 测试VLM：做一杯咖啡 {"On(Coffee,CoffeeTable)"}
3 测试VLM：做一杯咖啡放到吧台上 测试VLM：做一杯咖啡放到吧台上 {"On(Coffee,Bar)"}
4 测试VLM：做一杯咖啡放到水杯桌上并倒水 测试VLM：做一杯咖啡放到水杯桌上，再倒一杯水 测试VLM：做一杯咖啡放到水杯桌上并倒水 测试VLM：做一杯咖啡放到水杯桌上，再倒一杯水 {"On(Coffee,WaterTable)"} {"On(Coffee,WaterTable)","On(Water,WaterTable)"}
5 测试VLN：前往2号桌 测试VLN：前往2号桌 {"At(Robot,Table2)"}
6 测试AEM 测试AEM {"EnvExplored()"}
7 测试VLM：倒一杯水 测试VLM：倒一杯水 {"On(Water,WaterTable)"}

1
robowaiter/llm_client/data_tools/csv2dict.py
View File

@ -24,4 +24,3 @@ with open(csv_file_path, mode='r', encoding='gbk') as csv_file, \
json_str = json.dumps(output_dict, ensure_ascii=False)
# 将JSON字符串写入JSONL文件，并添加换行符
jsonl_file.write(json_str + '\n')
s

2
robowaiter/proto/camera.py
View File

@ -240,7 +240,7 @@ def get_semantic_map(camera, cur_objs, objs_name):
scene = Observe(0)
objs = scene.objects
img_data = get_camera([camera])
show_image(img_data, scene)
# show_image(img_data, scene)
objs_name = save_obj_info(img_data, objs_name)
for obj_name in list(objs_name):
for obj in objs:

BIN
robowaiter/proto/map_1.pkl Normal file
View File

Binary file not shown.

50
robowaiter/scene/scene.py
View File

@ -1,3 +1,4 @@
import sys
import time
import grpc
import numpy as np
@ -108,6 +109,7 @@ class Scene:
self.visited = set()
self.all_frontier_list = set()
self.semantic_map = semantic_map
self.auto_map = np.ones((800, 1550))
def reset(self):
@ -578,7 +580,7 @@ class Scene:
# v_list = [[0.0, 0.0]]
for walk_v in v_list:
walk_v = walk_v + [scene.rotation.Yaw - 90, 250, 60]
walk_v = walk_v + [scene.rotation.Yaw - 90, 250, 10]
print("walk_v", walk_v)
action = GrabSim_pb2.Action(scene=scene_id, action=GrabSim_pb2.Action.ActionType.WalkTo, values=walk_v)
scene = stub.Do(action)
@ -587,7 +589,7 @@ class Scene:
print(scene.info)
return cur_objs, objs_name_set
def isOutMap(self, pos, min_x=-350, max_x=600, min_y=-400, max_y=1450):
def isOutMap(self, pos, min_x=-200, max_x=600, min_y=-250, max_y=1300):
if pos[0] <= min_x or pos[0] >= max_x or pos[1] <= min_y or pos[1] >= max_y:
return True
return False
@ -598,11 +600,13 @@ class Scene:
'''
# x = round((x - self.min_x) / self.scale_ratio)
# y = round((y - self.min_y) / self.scale_ratio)
x = math.floor((x + 350) / 5)
y = math.floor((y + 400) / 5)
x = math.floor((x + 200))
y = math.floor((y + 250))
return x, y
def explore(self, map, cur_pos, explore_range):
def explore(self, map, explore_range):
scene = stub.Observe(GrabSim_pb2.SceneID(value=0))
cur_pos = [int(scene.location.X), int(scene.location.Y)]
for i in range(cur_pos[0] - explore_range, cur_pos[0] + explore_range + 1):
for j in range(cur_pos[1] - explore_range, cur_pos[1] + explore_range + 1):
if self.isOutMap((i, j)):
@ -610,6 +614,7 @@ class Scene:
x, y = self.real2map(i, j)
if map[x, y] == 0:
self.visited.add((i, j))
self.auto_map[x][y] = 0
for i in range(cur_pos[0] - explore_range, cur_pos[0] + explore_range + 1):
for j in range(cur_pos[1] - explore_range, cur_pos[1] + explore_range + 1):
if self.isOutMap((i, j)):
@ -618,11 +623,26 @@ class Scene:
if map[x, y] == 0:
if self.isNewFrontier((i, j), map):
self.all_frontier_list.add((i, j))
if len(self.all_frontier_list) <= 400:
if len(self.all_frontier_list) == 0:
free_list = list(self.visited)
free_array = np.array(free_list)
print(f"探索完成！以下是场景中可以到达的点：{free_array}；其余点均是障碍物不可达")
print(f"主动探索完成！以下是场景中可以到达的点：{free_array}；其余点均是障碍物不可达")
# 画地图: X行Y列，第一行在下面
plt.clf()
plt.imshow(self.auto_map, cmap='binary', alpha=0.5, origin='lower',
extent=(-250, 1300,
-200, 600))
plt.show()
print("已绘制完成地图！！！")
return None
# 画地图: X行Y列，第一行在下面
plt.imshow(self.auto_map, cmap='binary', alpha=0.5, origin='lower',
extent=(-250, 1300,
-200, 600))
plt.show()
print("已绘制部分地图！")
return self.getNearestFrontier(cur_pos, self.all_frontier_list)
def isNewFrontier(self, pos, map):
@ -630,12 +650,26 @@ class Scene:
for node in around_nodes:
x, y = self.real2map(node[0], node[1])
if node not in self.visited and map[x, y] == 0:
if not self.isOutMap((node[0], node[1])) and node not in self.visited and map[x, y] == 0:
return True
if (pos[0], pos[1]) in self.all_frontier_list:
self.all_frontier_list.remove((pos[0], pos[1]))
return False
def getDistance(self, pos1, pos2):
return math.sqrt((pos1[0] - pos2[0]) ** 2 + (pos1[1] - pos2[1]) ** 2)
def getNearestFrontier(self, cur_pos, frontiers):
dis_min = sys.maxsize
frontier_best = None
for frontier in frontiers:
dis = self.getDistance(frontier, cur_pos)
if dis <= dis_min:
dis_min = dis
frontier_best = frontier
return frontier_best
def cal_distance_to_robot(self,objx,objy,objz):
scene = self.status

21
robowaiter/scene/tasks/AEM.py
View File

@ -2,6 +2,7 @@
环境主动探索和记忆
要求输出探索结果（语义地图）对环境重点信息记忆。生成环境的语义拓扑地图，和不少于10个环境物品的识别和位置记忆，可以是图片或者文字或者格式化数据。
"""
import pickle
from robowaiter.scene.scene import Scene
class SceneAEM(Scene):
@ -12,9 +13,25 @@ class SceneAEM(Scene):
pass
def _run(self):
cur_objs = []
objs_name_set = set()
file_name = '../../proto/map_1.pkl'
if os.path.exists(file_name):
with open(file_name, 'rb') as file:
map = pickle.load(file)
print('------------ 自主探索 ------------')
cur_objs = self.semantic_map.navigation_move(cur_objs, 0, 11)
print("物品列表如下：")
# cur_objs = self.semantic_map.navigation_move(cur_objs, 0, 11)
# print("物品列表如下：")
# print(cur_objs)
# cur_pos = [int(scene.location.X), int(scene.location.Y)]
# print(reachable([237,490]))
# navigation_move([[237,490]], i, map_id)
# navigation_test(i,map_id)
while True:
goal = self.explore(map, 120) # cur_pos 指的是当前机器人的位置，场景中应该也有接口可以获取
if goal is None:
break
cur_objs, objs_name_set = self.navigation_move(cur_objs, objs_name_set, [[goal[0], goal[1]]], 0, 11)
print("------------物品信息--------------")
print(cur_objs)
pass

5
robowaiter/scene/tasks/VLM.py → robowaiter/scene/tasks/VLM/VLM.py
View File

@ -22,7 +22,10 @@ class SceneVLM(Scene):
# (5, self.create_chat_event("测试VLM：把冰红茶放到Table2")),
# (5, self.create_chat_event("测试VLM：关大厅灯"))
# (5, self.create_chat_event("测试VLM：做一杯咖啡放到吧台上")),
(5, self.create_chat_event("测试VLM：做一杯咖啡放到水杯桌上并倒水")),
(5, self.create_chat_event("测试VLM：做一杯咖啡放到水杯桌上，再倒一杯水")),
(10, self.create_chat_event("测试VLM：关窗帘")),
# (5, self.create_chat_event("测试VLN：前往2号桌")),
]
def _reset(self):

123
robowaiter/scene/tasks/VLM/VLM2.py Normal file
View File

@ -0,0 +1,123 @@
"""
视觉语言操作
机器人根据指令人的指令调节空调，自主探索环境导航到目标点，通过手臂的运动规划能力操作空调，比如开关按钮、调温按钮、显示面板
"""
import time
from robowaiter.scene.scene import Scene
class SceneVLM(Scene):
def __init__(self, robot):
super().__init__(robot)
# 在这里加入场景中发生的事件， (事件发生的时间，事件函数)
self.event_list = [
# (5, self.create_chat_event("测试VLM：做一杯咖啡")),
# (5, self.create_chat_event("测试VLM：倒一杯水")),
# (5, self.create_chat_event("测试VLM：开空调")),
# (5, self.create_chat_event("测试VLM：关空调")),
# (5, self.create_chat_event("测试VLM：开大厅灯")),
# (5, self.create_chat_event("测试VLM：拖地")),
# (7, self.create_chat_event("测试VLM：擦桌子")),
# (5, self.create_chat_event("测试VLM：整理椅子")),
# (5, self.create_chat_event("测试VLM：把冰红茶放到Table2")),
# (5, self.create_chat_event("测试VLM：关大厅灯"))
# (5, self.create_chat_event("测试VLM：做一杯咖啡放到吧台上")),
# (5, self.create_chat_event("测试VLM：做一杯咖啡放到水杯桌上并倒水")),
# (8, self.create_chat_event("测试VLN：前往1号桌")),
]
def _reset(self):
# self.gen_obj(type=5)
# self.gen_obj(type=9)
# self.op_task_execute(op_type=16, obj_id=0)
# self.move_task_area(op_type=4)
pass
def _run(self, op_type=10):
# 一个行人从门口走到 吧台
# 打招呼需要什么
# 行人说 哪里有位置，想晒个太阳
# 带领行人去有太阳的地方
# 行人说 有点热
# 好的，这就去开空调
scene = self.add_walkers([[0, 0]])
self.control_walker(
[self.walker_control_generator(walkerID=1, autowalk=False, speed=50, X=100, Y=150, Yaw=0)])
cont = scene.walkers[0].name+":我有点热，能开个空调吗？"
self.control_robot_action(0,3,cont)
# 共17个操作
# "制作咖啡","倒水","夹点心","拖地","擦桌子","开筒灯","搬椅子", # 1-7
# "关筒灯","开大厅灯","关大厅灯","关闭窗帘","打开窗帘", # 8-12
# "调整空调开关","调高空调温度","调低空调温度", # 13-15
# "抓握物体","放置物体" # 16-17
# self.gen_obj()
# if op_type <=15:
# self.move_task_area(op_type)
# self.op_task_execute(op_type)
# if op_type == 16: # 16: 抓操作需要传入物品id
# self.move_task_area(op_type, obj_id=0)
# self.op_task_execute(op_type, obj_id=0)
# # 原始吧台处:[247.0, 520.0, 100.0], 空调开关旁吧台:[240.0, 40.0, 100.0], 水杯桌:[-70.0, 500.0, 107]
# # 桌子1:[-55.0, 0.0, 107],抹布桌:[340.0, 900.0, 99.0] # 桌子2:[-55.0, 150.0, 107],
# if op_type == 17: # 17: 放操作需要传入放置位置周围的可达区域
# pos = [240.0, 40.0, 100.0]
# self.move_task_area(op_type, release_pos=pos)
# self.op_task_execute(op_type, release_pos=pos) # [325.0, 860.0, 100]
# 流程测试
# 抓握放置:抓吧台前生成的酸奶，放到抹布桌上
self.gen_obj()
# self.move_task_area(16, obj_id=0)
# self.op_task_execute(16, obj_id=0)
# pos = [340.0, 900.0, 99.0]
# self.move_task_area(17, release_pos=pos)
# self.op_task_execute(17, release_pos=pos)
#
# # 做咖啡:做完的咖啡放到水杯桌上
# self.move_task_area(1)
# self.op_task_execute(1)
#
# self.find_obj("CoffeeCup")
#
# self.move_task_area(16, obj_id=275)
# self.op_task_execute(16, obj_id=275)
# pos = [-70.0, 500.0, 107]
# self.move_task_area(17, release_pos=pos)
# self.op_task_execute(17, release_pos=pos)
#
# # 倒水:倒完的水放到旁边桌子上
# self.move_task_area(2)
# self.op_task_execute(2)
#
# self.move_task_area(16, obj_id=190)
# self.op_task_execute(16, obj_id=190)
# pos = [-55.0, 0.0, 107]
# self.move_task_area(17, release_pos=pos)
# self.op_task_execute(17, release_pos=pos)
# self.test_yaw()
pass
def _step(self):
pass
if __name__ == '__main__':
import os
from robowaiter.robot.robot import Robot
robot = Robot()
# create task
task = SceneVLM(robot)
task.reset()
task.run()

0
robowaiter/scene/tasks/VLM/__init__.py Normal file
View File

7
sub_task.ptml
View File

@ -1,5 +1,10 @@
selector{
cond On(Coffee,CoffeeTable)
sequence{
cond On(Coffee,CoffeeTable)
cond On(Coffee,CoffeeTable)
}
sequence{
cond Holding(Nothing)
act Make(Coffee)